1 2013-08-23 Jan Hubicka <jh@suse.cz>
3 * cgraph.c (cgraph_turn_edge_to_speculative):
6 2013-08-23 Jan Hubicka <jh@suse.cz>
8 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
9 (tree_decl_with_vis): Add FINAL field.
11 2013-08-23 Jeff Law <law@redhat.com>
13 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
14 the debugging dump when the expression is fully redundant.
16 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
18 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
19 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
20 * pretty-print.c (pp_formatted_text_data): Likewise.
21 (pp_write_text_to_stream): Likewise.
22 (pp_write_text_as_dot_label_to_stream): Likewise.
23 (pp_append_r): Likewise.
24 (pp_format): Likewise.
26 (pp_clear_output_area): Likewise.
27 (pp_append_text): Likewise.
28 (pp_formatted_text): Likewise.
29 (pp_remaining_character_count_for_line): Likewise.
30 (pp_newline): Likewise.
31 (pp_character): Likewise.
32 (output_buffer::~output_buffer): Define.
33 (pretty_printer::~pretty_printer): Destruct output buffer.
34 * pretty-print.h (output_buffer::~output_buffer): Declare.
35 (pretty_printer::~pretty_printer): Declare virtual.
37 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
40 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
41 HOST_WIDE_INT_M1U): New macros.
42 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
43 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
44 unsigned -1 for lshift.
45 * cse.c (cse_insn): Likewise.
46 * double-int.c (rshift_double, lshift_double): Likewise.
47 * builtins.c (fold_builtin_bitop): Likewise.
48 * combine.c (force_to_mode): Likewise.
49 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
50 * simplify-rtx.c (simplify_const_unary_operation,
51 simplify_const_binary_operation): Likewise.
52 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
53 check_va_list_escapes): Likewise.
54 * rtlanal.c (nonzero_bits1): Likewise.
55 * expmed.c (expand_smod_pow2): Likewise.
56 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
58 2013-08-23 Jan Hubicka <jh@suse.cz>
60 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
61 as having address taken.
63 2013-08-23 Jan Hubicka <jh@suse.cz>
65 * ipa-utils.h (method_class_type): Declare.
66 * ipa-devirt.c (method_class_type): Export.
68 * cgraphunit.c (analyze_functions): Do basic devirtualization;
69 do not walk base classes of anonymous types.
71 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
73 PR rtl-optimization/58220
75 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
76 handle SEQUENCE insns properly.
78 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
80 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
82 (pp_newline_and_indent): Likewise.
83 (pp_separate_with): Likewise.
84 * pretty-print.c (pp_newline_and_flush): Define.
85 (pp_newline_and_indent): Likewise.
86 (pp_separate_with): Likewise.
88 2013-08-23 Jakub Jelinek <jakub@redhat.com>
91 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
92 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
94 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
96 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
97 * gcc/config/i386/i386.md (*movti_internal): Use
98 predicate to determine if EVEX is needed.
99 (*movsi_internal): Ditto.
100 (*movdf_internal): Ditto.
101 (*movsf_internal): Ditto.
102 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
104 2013-08-23 Jakub Jelinek <jakub@redhat.com>
106 PR tree-optimization/58209
107 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
108 (find_tail_calls): Give up for pointer result types if m is non-NULL.
109 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
110 emit POINTER_PLUS_EXPR.
111 (create_tailcall_accumulator): For pointer result type accumulate in
114 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
116 * configure.ac: Add backslashes missing from the last change.
117 * configure: Regenerate.
119 2013-08-22 Jan Hubicka <jh@susue.cz>
121 * ipa.c (function_and_variable_visibility): First remember function
122 was global and then make it local.
124 2013-08-22 Julian Brown <julian@codesourcery.com>
126 * configure.ac: Add aarch64 to list of arches which use "nop" in
128 * configure: Regenerate.
130 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
132 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
133 gnu_libc_has_function.
134 * config/s390/tpf.h: Likewise.
136 2013-08-22 Jan Hubicka <jh@susue.cz>
138 * timevar.c (validate_phases): Add cast.
140 2013-08-22 Jan Hubicka <jh@susue.cz>
142 * timevar.c (validate_phases): Use size_t for memory.
143 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
145 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
147 * pretty-print.h (output_buffer::output_buffer): Declare.
148 (pretty_printer::pretty_printer): Likewise.
149 (pp_construct): Remove.
150 * pretty-print.c (output_buffer::output_buffer): Define.
151 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
152 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
153 (print_gimple_expr): Likewise.
154 (print_gimple_seq): Likewise.
155 (gimple_dump_bb): Likewise.
156 * sched-vis.c (dump_value_slim): Likewise.
157 (dump_insn_slim): Likewise.
158 (dump_rtl_slim): Likewise.
159 (str_pattern_slim): Likewise.
160 * tree-mudflap.c (mf_varname_tree): Likewise.
161 * graph.c (print_graph_cfg): Likewise.
162 (start_graph_dump): Likewise.
163 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
165 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
166 pretty printer initialization.
167 * coretypes.h (diagnostic_context): Remove superflous type alias
169 (pretty_printer): Likewise. Declare directly as a class.
170 (pretty_print_info): Remove declaration as class.
171 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
172 and pp_clear_output_area.
173 (asan_add_global): Likewise.
175 2013-08-22 Jan Hubicka <jh@suse.cz>
177 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
178 * ipa-utils.h (update_type_inheritance_graph): Declare.
179 (possible_polymorphic_call_target_p): Declare.
180 (possible_polymorphic_call_target_p): New.
181 * ipa-devirt.c: Update toplevel comments.
182 (cached_polymorphic_call_targets): Move up.
183 (odr_type_d): Move ID down.
184 (polymorphic_type_binfo_p): Update comment.
185 (odr_hasher::remove): Likewise;
186 (get_odr_type): Set anonymous_namespace.
187 (dump_odr_type): Dump it.
188 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
189 (maybe_record_node): Record node in cached_polymorphic_call_targets.
190 (record_binfo): Add comment.
191 (free_polymorphic_call_targets_hash): Do not ICE when cache is not built.
192 (devirt_node_removal_hook): Do not iCE when cache is freed.
193 (possible_polymorphic_call_target_p): New predicate.
194 (update_type_inheritance_graph): New function.
196 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
197 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
198 Sergey Lega <sergey.s.lega@intel.com>
199 Anna Tikhonova <anna.tikhonova@intel.com>
200 Ilya Tocar <ilya.tocar@intel.com>
201 Andrey Turetskiy <andrey.turetskiy@intel.com>
202 Ilya Verbin <ilya.verbin@intel.com>
203 Kirill Yukhin <kirill.yukhin@intel.com>
204 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
206 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
207 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
208 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
209 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
210 (OPTION_MASK_ISA_AVX2_UNSET): Update.
211 (OPTION_MASK_ISA_AVX512F_UNSET): New.
212 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
213 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
214 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
215 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
216 OPT_mavx512pf, OPT_mavx512er cases.
217 * config/i386/constraints.md (v): New constraint.
218 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
219 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
221 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
222 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
223 * config/i386/i386-c.c (ix86_target_macros_internal):
224 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
226 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
227 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
228 * config/i386/i386.c (regclass_map, dbx_register_map)
229 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
230 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
231 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
233 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
234 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
235 -mavx512pf options. Fix formatting.
236 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
237 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
238 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
239 -mavx512cd, -mavx512pf options.
240 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
241 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
242 (ix86_preferred_output_reload_class): Replace SSE_REGS with
244 (ix86_hard_regno_mode_ok): Support 512-bit registers.
245 (ix86_set_reg_reg_cost): Ditto.
246 (x86_order_regs_for_local_alloc): Ditto.
247 (MAX_VECT_LEN): Extend to 64-byte.
248 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
249 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
250 (TARGET_AVX512ER, TARGET_AVX512CD): New.
251 (BIGGEST_ALIGNMENT): Extend to 512-bits.
252 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
253 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
254 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
255 (SSE_REG_MODE_P): Support new modes.
256 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
257 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
258 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
259 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
260 (REG_CLASS_CONTENTS): Add new registers.
261 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
262 (EXT_REX_SSE_REGNO_P): New.
263 (HI_REGISTER_NAMES): Add new registers.
264 * config/i386/i386.md: Define constants for new registers.
265 (mode): Add new 512-bit modes.
266 (prefix): Support evex prefix.
267 (isa): Support avx512f, noavx512f, fma_avx512f.
268 (ssemodesuffix): Add new 512-bit modes.
270 (*movxi_internal_avx512f): Ditto.
271 (*movdi_internal): Replace constraint "x" with the new constraint "v".
273 (*movsi_internal): Likewise.
274 (*movdf_internal): Likewise.
275 (*movsf_internal): Likewise.
276 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
277 (<code><mode>3): Likewise.
278 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd): New.
279 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
280 with the new constraint "v".
281 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
283 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
284 with the new constraint "v".
285 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
286 (<sse2>_storedqu<avxsizesuffix>): Likewise.
287 (*<plusminus_insn><mode>3): Likewise.
288 (<sse>_vm<plusminus_insn><mode>3): Likewise.
289 (*mul<mode>3): Likewise.
290 (<sse>_vmmul<mode>3): Likewise.
291 (<sse>_div<mode>3): Likewise.
292 (<sse>_vmdiv<mode>3): Likewise.
293 (<sse>_sqrt<mode>2): Likewise.
294 (<sse>_vmsqrt<mode>2): Likewise.
295 (*<code><mode>3_finite): Likewise.
296 (*<code><mode>3) <smaxmin>: Likewise.
297 (<sse>_vm<code><mode>3): Likewise.
298 (*<code><mode>3) <any_logic>: Likewise.
299 (*fma_fmadd_<mode>): Likewise.
300 (*fma_fmsub_<mode>): Likewise.
301 (*fma_fnmadd_<mode>): Likewise.
302 (*fma_fnmsub_<mode>): Likewise.
303 (*fma_fmaddsub_<mode>): Likewise.
304 (*fma_fmsubadd_<mode>): Likewise.
305 (*fmai_fmadd_<mode>): Likewise.
306 (*fmai_fmsub_<mode>): Likewise.
307 (*fmai_fnmadd_<mode>): Likewise.
308 (*fmai_fnmsub_<mode>): Likewise.
309 (sse_cvtsi2ss): Likewise.
310 (sse_cvtsi2ssq): Likewise.
311 (sse_cvtss2si): Likewise.
312 (sse_cvtss2si_2): Likewise.
313 (sse_cvtss2siq): Likewise.
314 (sse_cvtss2siq_2): Likewise.
315 (sse_cvttss2si): Likewise.
316 (sse_cvtss2siq_2): Likewise.
317 (float<sseintvecmodelower><mode>2): Likewise.
318 (sse2_cvtsd2si_2): Likewise.
319 (sse2_cvtsd2siq_2): Likewise.
320 (*<plusminus_insn><mode>3): Likewise.
321 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
322 (*<sse4_1_avx2>_mul<mode>3): Likewise.
323 (ashr<mode>3): Likewise.
324 (<shift_insn><mode>3): Likewise.
325 (avx2_<code><mode>3): Likewise.
326 (*avx2_<code><mode>3): Likewise.
327 (*andnot<mode>3): Likewise.
328 (*<code><mode>3) <any_logic>: Likewise.
329 (abs<mode>2): Likewise.
330 (avx2_permvar<mode>): Likewise.
331 (avx2_perm<mode>_1): Likewise.
332 (*avx_vpermilp<mode>): Likewise.
333 (avx_vpermilvar<mode>3): Likewise.
334 (avx2_ashrv<mode>): Likewise.
335 (avx2_<shift_insn>v<mode>): Likewise.
336 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
338 * doc/rtl.texi: Document XImode.
340 2013-08-21 Jeff Law <law@redhat.com>
342 * tree-flow.h (register_jump_thread): Pass vector of edges
343 instead of each important edge.
344 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
345 thread path into a vector and pass that to register_jump_thread.
346 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
347 passed in edge vector to the current 3-edge form.
350 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
352 * dce.c (fini_dce): Call df_analyze again just in case
353 delete_unmarked_insns removed anything.
355 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
357 * reload.h (struct reg_equivs): Rename to ..
358 (struct reg_equivs_s): .. this.
360 2013-08-20 Martin Liska <marxin.liska@gmail.com>
362 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
364 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
366 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest
369 2013-08-21 Jeff Law <law@redhat.com>
371 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
372 simplify assignments too. If the RHS collapses to a singleton
373 range, then return the value for the range.
375 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
377 * config/i386/sse.md (V16): Rename to...
379 (mov<mode>): Update iterator name.
380 (*mov<mode>_internal): Ditto.
381 (push<mode>1): Ditto.
382 (movmisalign<mode>): Ditto.
384 2013-08-20 Jan Hubicka <jh@suse.cz>
387 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
388 entry for direct edges.
389 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
391 2013-08-20 David Malcolm <dmalcolm@redhat.com>
393 Revert my last two changes, r201865 and r201864:
396 2013-08-20 David Malcolm <dmalcolm@redhat.com>
398 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
399 instances can own GC refs.
401 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
402 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
403 (gcc::context::gt_pch_nx): Likewise.
404 (gcc::context::gt_pch_nx): Likewise.
405 * ggc.h (gt_ggc_mx <T>): New.
406 (gt_pch_nx_with_op <T>): New.
407 (gt_pch_nx <T>): New.
408 * passes.c (opt_pass::gt_ggc_mx): New.
409 (opt_pass::gt_pch_nx): New.
410 (opt_pass::gt_pch_nx_with_op): New.
411 (pass_manager::gt_ggc_mx): New.
412 (pass_manager::gt_pch_nx): New.
413 (pass_manager::gt_pch_nx_with_op): New.
414 (pass_manager::operator new): Use
415 ggc_internal_cleared_alloc_stat rather than xcalloc.
416 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
417 (pass_manager::gt_ggc_mx): New.
418 (pass_manager::gt_pch_nx): New.
419 (pass_manager::gt_pch_nx_with_op): New.
420 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
421 (opt_pass::operator new): New.
422 (opt_pass::gt_ggc_mx): New.
423 (opt_pass::gt_pch_nx): New.
424 (opt_pass::gt_pch_nx_with_op): New.
427 2013-08-20 David Malcolm <dmalcolm@redhat.com>
429 * Makefile.in (GTFILES): Add context.h.
430 * context.c (gcc::context::operator new): New.
431 (gcc::context::gt_ggc_mx): New.
432 (gcc::context::gt_pch_nx): New.
433 (gcc::context::gt_pch_nx): New.
434 * context.h (gcc::context): Add GTY((user)) marking.
435 (gcc::context::operator new): New.
436 (gcc::context::gt_ggc_mx): New.
437 (gcc::context::gt_pch_nx): New.
438 (gcc::context::gt_pch_nx): New.
439 (g): Add GTY marking.
440 (gt_ggc_mx (gcc::context *)): New.
441 (gt_pch_nx (gcc::context *)): New.
442 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
444 * gengtype.c (open_base_files) <ifiles>: Add context.h.
446 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
448 * dce.c (fini_dce): Call df_analyze again just in case
449 delete_unmarked_insns removed anything.
451 2013-08-20 Teresa Johnson <tejohnson@google.com>
453 PR rtl-optimizations/57451
454 * final.c (reemit_insn_block_notes): Prevent lexical blocks
455 from crossing split section boundaries.
457 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
459 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
460 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
461 with MULTLIB_DEFAULTS.
463 2013-08-20 Nick Clifton <nickc@redhat.com>
465 * target.def (narrow_volatile_bitfield): Note that the default
466 value is false, not !TARGET_STRICT_ALIGN.
467 * doc/tm.texi: Regenerate.
469 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
471 Fix LIB_SPEC for systems without libpthread.
473 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
474 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
476 * config/i386/linux-common.h: Likewise.
477 * config/mips/linux-common.h: Likewise.
479 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
481 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
484 2013-08-20 David Malcolm <dmalcolm@redhat.com>
486 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
487 instances can own GC refs.
489 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
490 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
491 (gcc::context::gt_pch_nx): Likewise.
492 (gcc::context::gt_pch_nx): Likewise.
493 * ggc.h (gt_ggc_mx <T>): New.
494 (gt_pch_nx_with_op <T>): New.
495 (gt_pch_nx <T>): New.
496 * passes.c (opt_pass::gt_ggc_mx): New.
497 (opt_pass::gt_pch_nx): New.
498 (opt_pass::gt_pch_nx_with_op): New.
499 (pass_manager::gt_ggc_mx): New.
500 (pass_manager::gt_pch_nx): New.
501 (pass_manager::gt_pch_nx_with_op): New.
502 (pass_manager::operator new): Use
503 ggc_internal_cleared_alloc_stat rather than xcalloc.
504 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
505 (pass_manager::gt_ggc_mx): New.
506 (pass_manager::gt_pch_nx): New.
507 (pass_manager::gt_pch_nx_with_op): New.
508 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
509 (opt_pass::operator new): New.
510 (opt_pass::gt_ggc_mx): New.
511 (opt_pass::gt_pch_nx): New.
512 (opt_pass::gt_pch_nx_with_op): New.
514 2013-08-20 David Malcolm <dmalcolm@redhat.com>
516 * Makefile.in (GTFILES): Add context.h.
517 * context.c (gcc::context::operator new): New.
518 (gcc::context::gt_ggc_mx): New.
519 (gcc::context::gt_pch_nx): New.
520 (gcc::context::gt_pch_nx): New.
521 * context.h (gcc::context): Add GTY((user)) marking.
522 (gcc::context::operator new): New.
523 (gcc::context::gt_ggc_mx): New.
524 (gcc::context::gt_pch_nx): New.
525 (gcc::context::gt_pch_nx): New.
526 (g): Add GTY marking.
527 (gt_ggc_mx (gcc::context *)): New.
528 (gt_pch_nx (gcc::context *)): New.
529 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
531 * gengtype.c (open_base_files) <ifiles>: Add context.h.
533 2013-08-20 Alan Modra <amodra@gmail.com>
536 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
537 (rs6000_emit_epilogue): Likewise.
539 2013-08-19 Dehao Chen (dehao@google.com)
541 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
543 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
544 Jakub Jelinek <jakub@redhat.com>
546 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
547 (BUILT_IN_FABSD64): Likewise.
548 (BUILT_IN_FABSD128): Likewise.
549 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
550 (fold_builtin_1): Likewise.
551 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping destination
553 (*abstd2_fpr): Likewise.
554 (*nabstd2_fpr): Likewise.
556 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
558 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
561 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
563 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
566 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
568 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
569 * builtins.c (default_libc_has_function): New.
570 (gnu_libc_has_function): Ditto.
571 (no_c99_libc_has_function): Ditto.
572 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
573 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
574 (fold_builtin_sincos): Likewise.
575 (fold_builtin_cexp): Likewise.
576 * builtins.def (DEF_C94_BUILTIN): Likewise.
577 (DEF_C99_BUILTIN): Likewise.
578 (DEF_C99_C90RES_BUILTIN): Likewise.
579 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
580 definitions to using this define.
581 * config/darwin-protos.h (darwin_libc_has_function): New.
582 * config/darwin.c: (darwin_libc_has_function: Ditto.
583 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
584 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
585 * config/darwin.h: Ditto.
586 * config/elfos.h: Ditto.
587 * config/freebsd.h: Ditto.
588 * config/i386/cygming.h: Ditto.
589 * config/i386/djgpp.h: Ditto.
590 * config/i386/i386-interix.h: Ditto.
591 * config/microblaze/microblaze.h: Ditto.
592 * config/mmix/mmix.h: Ditto.
593 * config/gnu-user.h: Ditto.
594 * config/ia64/hpux.h: Ditto.
595 * config/pa/pa-hpux.h: Ditto.
596 * config/pdp11/pdp11.h: Ditto.
597 * config/picochip/picochip.h: Ditto.
598 * config/linux.h: Ditto.
599 * config/netbsd.h: Ditto.
600 * config/openbsd.h: Ditto.
601 * config/rs6000/aix43.h: Ditto.
602 * config/rs6000/aix51.h: Ditto.
603 * config/rs6000/aix52.h: Ditto.
604 * config/rs6000/aix53.h: Ditto.
605 * config/rs6000/aix61.h: Ditto.
606 * config/rs6000/darwin.h: Ditto.
607 * config/rs6000/linux.h: Ditto.
608 * config/rs6000/linux64.h: Ditto.
609 * config/s390/tpf.h: Ditto.
610 * config/sol2-10.h: Ditto.
611 * config/sol2.h: Ditto.
612 * config/vms/vms.h: Ditto.
613 * config/vxworks.h: Ditto.
614 * config/linux-android.c (linux_android_libc_has_function):
615 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
616 * config/linux-protos.h (linux_android_libc_has_function):
618 * config/i386/i386.c (ix86_libc_has_function): New.
619 * config/i386/i386-protos.h
620 (ix86_libc_has_function): New declaration.
621 * config/i386/i386.md
622 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
623 ("isinf<mode>2): Likewise.
624 * convert.c (convert_to_integer): Using new target hook
625 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
626 TARGET_C99_FUNCTIONS.
627 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
628 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
629 * coretypes.h (function_class): New enum for different
630 classes of functions.
631 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
632 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
633 (TARGET_HAS_SINCOS): Likewise.
634 (TARGET_LIBC_HAS_FUNCTION): New.
635 * doc/tm.texi: Regenerated.
636 * targhooks.h (default_libc_has_function): New declaration.
637 (no_c99_libc_has_function): Ditto.
638 (gnu_libc_has_function): Ditto.
639 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
640 and TARGET_HAS_SINCOS.
642 2013-08-18 Jan Hubicka <jh@suse.cz>
644 * Makeifle-in (ipa-devirt.o): New.
645 (GTFILES): Add ipa-utils.h and ipa-devirt.c
646 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
647 (analyze_functions): Look into possible targets of polymorphic call.
648 * dumpfile.c (dump_files): Add type-inheritance dump.
649 * dumpfile.h (TDI_inheritance): New.
650 * ipa-devirt.c: New file.
651 * ipa-utils.h (odr_type_d): Forward declare.
652 (odr_type): New type.
653 (build_type_inheritance_graph): Declare.
654 (possible_polymorphic_call_targets): Declare and introduce inline
655 variant when only edge is pased.
656 (dump_possible_polymorphic_call_targets): Likewise.
657 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
658 * tree.c (type_in_anonymous_namespace_p): Break out from ...
659 (types_same_for_odr): ... here.
660 * tree.h (type_in_anonymous_namespace_p): Declare.
662 2013-08-18 Jakub Jelinek <jakub@redhat.com>
664 PR tree-optimization/58006
665 * tree-parloops.c (take_address_of): Don't ICE if get_name
667 (eliminate_local_variables_stmt): Remove clobber stmts.
669 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
671 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
672 error message is issued for an alias to undefined symbol.
674 2013-08-18 Jan Hubicka <jh@suse.cz>
676 * cgraph.c (cgraph_create_indirect_edge): Discover
677 polymorphic calls and record basic info into indirect_info.
678 * gimple-fold.c (gimple_fold_call): When doing BINFO based
679 devirtualization, ignore objc function calls.
680 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
681 call with no parm index info.
682 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
683 * tree.c (virtual_method_call_p): New function.
684 * tree.h (virtual_method_call_p): Declare.
686 2013-08-16 Jan Hubicka <jh@suse.cz>
689 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
691 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
693 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
696 2013-08-16 David Malcolm <dmalcolm@redhat.com>
698 * gengtype.c (type_for_name): Add special-case support for
699 locating types within the "gcc::" namespace.
700 (open_base_files): Emit a "using namespace gcc" directive.
702 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
705 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
706 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
708 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
709 array instead of each individual operand as a separate argument.
710 (emit_fusion_gpr_load): Likewise.
711 (expand_fusion_gpr_load): Add new function declaration.
713 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
714 signature to have the operands passed as an array, instead of as
715 separate arguments. Allow ZERO_EXTEND to be in the memory
716 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
717 depend on the register live/dead flags when peepholes are run.
718 (expand_fusion_gpr_load): New function to be called from the
719 peephole2 pass, to change the register that addis sets to be the
721 (emit_fusion_gpr_load): Change the calling signature to have the
722 operands passed as an array, instead of as separate arguments.
723 Allow ZERO_EXTEND to be in the memory address, and also
724 SIGN_EXTEND if -mpower8-fusion-sign.
726 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
728 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
729 adjust the register addis loads up in the peephole2 pass. Do not
730 depend on the register live/dead state when the peephole pass is
733 2013-08-16 David Malcolm <dmalcolm@redhat.com>
735 * gengtype.c (create_user_defined_type): Ensure that the kind
736 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
737 declaration is seen before the GTY((user)) marking.
739 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
742 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
744 2013-08-16 Jan Hubicka <jh@suse.cz>
746 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
747 arugment expected_type.
748 (gimple_fold_call): Use it.
749 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
750 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
751 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
752 (try_make_edge_direct_virtual_call): Likewise.
753 * tree.c (obj_type_ref_class): New.
754 * tree.h (obj_type_ref_class): Use it.
756 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
758 * sched-vis.c (rtl_slim_pp_initialized): Remove.
759 (rtl_slim_pp): Likewise.
760 (init_rtl_slim_pretty_print): Likewise.
761 (dump_value_slim): Don't call it. Use local pretty printer.
762 (dump_insn_slim): Likewise.
763 (dump_rtl_slim): Likewise.
764 (str_pattern_slim): Likewise.
765 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
768 2013-08-16 Jakub Jelinek <jakub@redhat.com>
770 PR tree-optimization/58164
771 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
772 walk gimple_goto_dest of GIMPLE_GOTO.
774 PR tree-optimization/58165
775 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
776 bi_call must be the last stmt in a bb, don't split_block, instead
777 use fallthru edge from it and give up if there is none.
778 Release conds vector when returning early.
780 2013-08-14 Xinliang David Li <davidxl@google.com>
782 * config/i386/i386.c (ix86_option_override_internal):
783 Remove unused variable and field.
785 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
788 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
789 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
790 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
791 and Linux, correct BLKmode alignment when 128-bit alignment is
792 required and compatibility flag is not set.
793 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
794 for zero-size arguments when compatibility flag is not set.
796 2013-08-14 Jakub Jelinek <jakub@redhat.com>
798 PR tree-optimization/58145
799 * tree-sra.c (build_ref_for_offset): If prev_base has
800 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
802 2013-08-14 Xinliang David Li <davidxl@google.com>
804 * config/i386/i386.c (ix86_option_override_internal):
805 Fix uninitialized variable error.
807 2013-08-14 Xinliang David Li <davidxl@google.com>
809 * config/i386/i386.opt: Define two new options.
810 * config/i386/x86-tune.def: Add arch selector field in macros.
811 * config/i386/i386.h: Adjust macro definition.
812 * config/i386/i386.c (ix86_option_override_internal):
814 (parse_mtune_ctrl_str): New function.
815 (set_ix86_tune_features): New function.
816 (ix86_function_specific_restore): Call the new helper function.
818 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
820 PR rtl-optimization/57662
821 * sel-sched.c (code_motion_process_successors): When the current insn
822 is removed after the recursive traversal, break from the loop.
823 Add comments and debug printouts.
825 2013-08-14 Jakub Jelinek <jakub@redhat.com>
826 Alexandre Oliva <aoliva@redhat.com>
829 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
830 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
831 there also UNSPEC_PLTOFF.
833 2013-08-14 Marek Polacek <polacek@redhat.com>
835 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
836 NUM_CONDITIONS bit positions.
838 2013-08-13 Cary Coutant <ccoutant@google.com>
840 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
841 (attr_checksum): Hash vector contents instead of pointer.
842 (attr_checksum_ordered): Likewise.
844 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
846 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
847 when Pmode != word_mode. Add length_address attribute.
848 (sse3_monitor_<mode>): Merge from sse3_monitor and
849 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
850 Pmode != word_mode. Update insn length attribute.
851 * config/i386/i386.c (ix86_option_override_internal): Update
852 ix86_gen_monitor selection for merged sse3_monitor insn.
854 2013-08-13 Julian Brown <julian@codesourcery.com>
856 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
857 perform invalid legitimization on greater-than-word-size modes for
860 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
862 * ira.c (setup_class_translate_array): Use aclass instead of cl
863 for classes not fully covered by allocno classes.
865 2013-08-13 Jakub Jelinek <jakub@redhat.com>
867 PR tree-optimization/57661
868 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
869 * tree-inline.c (tree_function_versioning): Initialize it.
870 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
871 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
872 that is not being copied.
875 * asan.c (instrument_strlen_call): Fix typo in comment.
876 Use char * type even for the lhs of POINTER_PLUS_EXPR.
878 2013-08-13 Steve Ellcey <sellcey@mips.com>
880 * config/mips/mips.md (prefetch): Use lw instead of ld on
881 loongson in 32bit mode.
883 2013-08-13 Nick Clifton <nickc@redhat.com>
885 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
887 2013-08-13 Jan Hubicka <jh@suse.cz>
889 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
890 introduced edge; fix typo in sanity check.
891 (cgraph_resolve_speculation): Export; improve diagnostic.
892 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
893 speculation at type mismatch.
894 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
895 (cgraph_resolve_speculation): Declare.
896 (symtab_can_be_discarded): New function.
897 * value-prof.c (gimple_ic_transform): Remove actual transform code.
898 * ipa-inline-transform.c (speculation_removed): New global var.
899 (clone_inlined_nodes): See if speculation can be removed.
900 (inline_call): If speculations was removed, we growths may not match.
901 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
902 (speculation_useful_p): New function.
903 (resolve_noninline_speculation): New function.
904 (inline_small_functions): Resolve useless speculations.
905 * ipa-inline.h (speculation_useful_p): Declare
906 * ipa.c (can_replace_by_local_alias): Simplify.
907 (ipa_profile): Produce speculative calls in non-lto, too;
908 add simple cost model; produce local aliases.
910 2013-08-13 David Malcolm <dmalcolm@redhat.com>
912 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
915 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
917 * config/i386/i386.c (ix86_function_versions): Use error + inform.
919 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
921 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
922 iterator instead of X87MODEF.
924 2013-08-12 Perez Read <netfirewall@gmail.com>
927 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
928 operand 0 for intel asm alternative.
929 (*movabs<mode>_2): Ditto for operand 1.
931 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
933 * config/aarch64/arm_none.h
934 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
936 2013-08-12 Nick Clifton <nickc@redhat.com>
938 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
940 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
942 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
943 expand for QI/HImode operand to produce more effictive code for
944 unsigned char(short) --> float(double) conversion.
946 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
948 * doc/invoke.texi: Mention that -ftls-model does not force the final
951 2013-08-12 Marek Polacek <polacek@redhat.com>
952 Marc Glisse <marc.glisse@inria.fr>
954 PR tree-optimization/57980
955 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
956 when creating -1 constant.
958 2013-08-10 Jan Hubicka <jh@suse.cz>
960 Workaround binutils PR14342.
961 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
962 (gimple_init_edge_profiler): Likewise.
963 (gimple_gen_ic_func_profiler): Likewise.
965 2013-08-09 Jan Hubicka <jh@suse.cz>
967 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
969 2013-08-09 Xinliang David Li <davidxl@google.com>
971 * config/i386/stringop.def: New file.
972 * config/i386/stringop.opt: New file.
973 * config/i386/i386-opts.h: Include stringopt.def.
974 * config/i386/i386.opt: Include stringopt.opt.
975 * config/i386/i386.c (ix86_option_override_internal):
976 Override default size based stringop inline strategies with options.
977 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
980 2013-08-09 Jan Hubicka <jh@suse.cz>
982 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
984 2013-08-09 Jan Hubicka <jh@suse.cz>
986 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
988 (dump_cgraph_node): Dump profile-id.
989 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
990 and common_target_probability.
991 * lto-cgraph.c (lto_output_edge): Stream common targets.
992 (lto_output_node): Stream profile ids.
993 (input_node): Stream profile ids.
994 (input_edge): Stream common targets.
995 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
996 * ipa.c: Include value-prof.h
997 (ipa_profile_generate_summary): Turn indirect call statement histograms
999 (ipa_profile): Turn common targets into speculative edges.
1001 2013-08-09 Jan Hubicka <jh@suse.cz>
1003 * cgraph.h (cgraph_node): Add profile_id.
1004 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1005 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1006 (del_node_map): Update.
1007 (find_func_by_funcdef_no): Replace by ...
1008 (find_func_by_profile_id): ... this one.
1009 (gimple_ic_transform): Do not remove useful histograms when
1010 speculation is not done; dump info when indirect call removal
1012 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1013 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1014 (__gcov_indirect_call_profiler_v2): .. this one.
1015 * profile.h (init_node_map): Update.
1016 * coverage.c (coverage_compute_profile_id): New function.
1017 * coverage.h (coverage_compute_profile_id): Declare.
1018 * tree-profile.c (init_ic_make_global_vars): Make
1019 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1020 (gimple_init_edge_profiler): Update prototype of
1021 __gcov_indirect_call_profiler.
1022 (gimple_gen_ic_func_profiler): Simplify.
1023 (tree_profiling): Use init_node_map
1025 2013-08-09 Jan Hubicka <jh@suse.cz>
1027 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1028 non-speculative refs.
1029 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1030 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1031 (cgraph_set_call_stmt): Likewise.
1032 (cgraph_create_edge_1): Fix release checking compilatoin;
1034 (cgraph_free_edge): Free indirect info.
1035 (cgraph_turn_edge_to_speculative): New function.
1036 (cgraph_speculative_call_info): New function.
1037 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1038 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1039 (dump_cgraph_node): Dump speculation.
1040 (verify_edge_count_and_frequency): Accept speculative edges.
1041 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1042 (verify_cgraph_node): Handle speculation.
1043 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1044 (cgraph_set_call_stmt): Update prototype.
1045 (cgraph_make_edge_direct): Update prototype.
1046 (cgraph_speculative_call_info): Declare.
1047 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1048 to change; update call of ipa_find_references.
1049 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1050 lto_stmt_uid and speculative flags.
1051 (ipa_dump_references): Dump speculation.
1052 (ipa_clone_references): Clone speculative flag.
1053 (ipa_clone_referring): Likewise.
1054 (ipa_clone_ref): New function.
1055 (ipa_find_reference): Look into lto_stmt_uids
1056 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1057 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1058 (ipa_find_reference): Update declaration.
1059 (ipa_clone_ref): Declare.
1060 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1061 stream speculative flag.
1062 (lto_output_ref): Stream statements uids and speculation.
1063 (input_ref): Likewise.
1064 (input_edge): Stream speuclation.
1065 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1066 (cgraph_set_call_stmt_including_clones): Handle speculation.
1067 * ipa-inline.c (heap_edge_removal_hook): New function.
1068 (inline_small_functions): Register it.
1069 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1070 also initialize refs.
1071 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1073 (try_make_edge_direct_simple_call): Likewise.
1074 (try_make_edge_direct_simple_call): Likewise.
1075 (update_indirect_edges_after_inlining): Likewise.
1076 (remove_described_reference): Look proper lto_stmt_uid.
1077 (propagate_controlled_uses): Likewise.
1078 (propagate_controlled_uses): Liekwise.
1079 * tree-inline.c (copy_bb): Copy speculative edges.
1080 (redirect_all_calls): New function.
1081 (copy_cfg_body): Do redirection after loop info is updated.
1082 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1084 2013-08-09 Jan Hubicka <jh@suse.cz>
1086 * lto-streamer-out.c (output_function): Renumber PHIs.
1087 * lto-streamer-in.c (input_function): Likewise.
1089 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1091 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1092 (get_lane_unsigned): Likewise.
1093 (dup_lane_scalar): Likewise.
1094 (get_lane): enable for VALL.
1095 * config/aarch64/aarch64-simd.md
1096 (aarch64_dup_lane_scalar<mode>): Remove.
1097 (aarch64_get_lane_signed<mode>): Likewise.
1098 (aarch64_get_lane_unsigned<mode>): Likewise.
1099 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1100 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1101 (aarch64_get_lane<mode>): Enable for all vector modes.
1102 (aarch64_get_lanedi): Remove misleading constraints.
1103 * config/aarch64/arm_neon.h
1104 (__aarch64_vget_lane_any): Define.
1105 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1106 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1107 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1108 * config/aarch64/iterators.md (VDQQH): New.
1112 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1114 * configure.ac: Add GAS check for LEON instructions on SPARC.
1115 * configure: Regenerate.
1116 * config.in: Likewise.
1117 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1119 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1120 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1122 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1123 (AS_LEON_FLAG): New macro.
1124 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1125 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1126 Deal with LEON and LEON3 for the memory model.
1127 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1128 (atomic_compare_and_swap<mode>_1): Likewise.
1129 (*atomic_compare_and_swap<mode>_1): Likewise.
1131 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1133 * config/arm/neon.md (vcond): Fix floating-point vector
1134 comparisons against 0.
1136 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1138 * lra-constraints.c (emit_spill_move): Remove assert.
1139 (process_alt_operands): Add more debugging
1140 output. Increase reject for spilling into memory. Decrease
1141 reject for reloading scratch.
1142 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1144 2013-08-08 Steve Ellcey <sellcey@mips.com>
1146 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1147 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1148 micromips incompatible. Add nan2008.
1149 (MULTILIB_DIRNAMES): Add nan2008.
1150 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1151 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1152 and micromips incompatible. Add nan2008.
1153 (MULTILIB_DIRNAMES): Add nan2008.
1154 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1156 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1158 PR rtl-optimization/58079
1159 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1160 simplify_comparison has widened a comparison with an integer.
1162 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1164 * config/arm/neon.md (movmisalign<mode>): Disable when we
1165 don't allow unaligned accesses.
1166 (*movmisalign<mode>_neon_store): Likewise.
1167 (*movmisalign<mode>_neon_load): Likewise.
1168 (*movmisalign<mode>_neon_store): Likewise.
1169 (*movmisalign<mode>_neon_load): Likewise.
1171 2013-08-08 Jan Hubicka <jh@suse.cz>
1173 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1174 (make_pass_rebuild_cgraph_edges): Also clear references.
1175 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1176 * ipa-inline-transform.c (inline_transform): Remove all references
1178 * cgraphunit.c (expand_function): Remove all references after
1180 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1181 (ipa_find_reference): Rewrite to iterator.
1182 (remove_stmt_references): Likewise.
1183 (ipa_clear_stmts_in_references): New function.
1184 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1185 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1187 * ipa-split.c (split_function): Remove references in split function.
1189 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1192 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1193 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1195 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1198 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
1199 suggested mode for the assignment isn't compatible with the
1202 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1205 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
1207 2013-08-07 Xinliang David Li <davidxl@google.com>
1209 * config/i386/i386.opt: New option -mtune-ctrl=.
1210 * config/i386/x86-tune.def: New file.
1211 * config/i386/i386.h: include x86-tune.def.
1212 * config/i386/i386.c (ix86_option_override_internal):
1213 Parsing -mtune-ctrl= option and set tune features.
1215 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
1218 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
1219 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
1222 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1224 * rtl.h (update_alignments): Declare.
1225 * final.c (grow_label_align): New function extracted from...
1226 (shorten_branches): ...here. Call it.
1227 (update_alignments): New function.
1228 * reorg.c (sibling_labels): New variable.
1229 (get_label_before): Add SIBLING parameter. If it is non-zero, push
1230 the new label along with it onto the sibling_labels vector.
1231 (fill_simple_delay_slots): Adjust call to get_label_before.
1232 (fill_slots_from_thread): Likewise.
1233 (relax_delay_slots): Likewise.
1234 (make_return_insns): Likewise.
1235 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
1237 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1239 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
1240 document its semantics.
1241 (diagnostic_report_diagnostic): Adjust accordingly.
1243 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1245 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
1246 (sparc_option_override): ...and port to new C++ pass API.
1247 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
1249 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
1251 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
1253 2013-08-06 Caroline Tice <cmtice@google.com>
1255 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
1256 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
1257 * tree-pass.h: Add pass_vtable_verify.
1258 * varasm.c (assemble_variable): Add code to properly set the comdat
1259 section and name for the .vtable_map_vars section.
1260 (assemble_vtyv_preinit_initializer): New function.
1261 (default_sectin_type_flags): Make sure .vtable_map_vars section has
1263 * output.h: Add function decl for assemble_vtv_preinit_initializer.
1264 * vtable-verify.c: New file.
1265 * vtable-verify.h: New file.
1266 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
1267 initialiation levels.
1268 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
1269 * passes.def: Insert pass_vtable_verify.
1270 * aclocal.m4: Reorder includes.
1271 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
1272 -fvtv-counts options.
1273 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
1274 as appropriate, if -fvtable-verify=... is used.
1275 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
1276 -fvtable-verify=... is used.
1277 * Makefile.in (OBJS): Add vtable-verify.o to list.
1278 (vtable-verify.o): Add new build rule.
1279 (GTFILES): Add vtable-verify.c to list.
1280 * common.opt (fvtable-verify=): New flag.
1281 (vtv_priority): Values for fvtable-verify= flag.
1282 (fvtv-counts): New flag.
1283 (fvtv-debug): New flag.
1284 * tree.h (save_vtable_map_decl): New extern function decl.
1286 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1288 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
1289 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
1290 (pass_data_rl78_devirt): ...new pass_data instance and...
1291 (make_pass_rl78_devirt): ...new function.
1292 (rl78_asm_file_start): Port pass registration to new C++ API.
1294 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1296 * coretypes.h (rtl_opt_pass): Add.
1297 (gcc::context): Add.
1298 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
1299 (epiphany_init): Port to new C++ pass API.
1300 (epiphany_optimize_mode_switching): Likewise.
1301 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
1302 (pass_manager::get_pass_mode_switching): New.
1303 (pass_manager::get_pass_peephole2): New.
1304 * mode-switching.c (pass_mode_switching): Add clone method.
1305 * recog.c (pass_peephole2): Add clone method.
1306 (pass_split_all_insns): Add clone method.
1308 2013-08-06 David Malcolm <dmalcolm@redhat.com>
1310 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
1311 (mips_option_override): ...here, porting to new C++ API for passes.
1313 2013-08-06 Jan Hubicka <jh@suse.cz>
1315 * cgraph.c (cgraph_get_body): New function based on lto.c
1317 * cgraph.h (cgraph_get_body): Declare.
1318 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
1320 * cgraphunit.c (expand_function): Get body prior expanding.
1321 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
1322 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
1324 * passes.c (do_per_function_toporder): Get body.
1325 * tree-inline.c (expand_call_inline): Get body prior inlining it.
1326 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
1328 2013-08-06 Martin Jambor <mjambor@suse.cz>
1331 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
1332 is not re-finalized. Rename second parameter to no_collect.
1334 2013-08-06 Martin Jambor <mjambor@suse.cz>
1337 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
1338 MEM_REF has proper alignment information.
1340 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
1343 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
1345 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
1346 rtx (*) (rtx, ...) with insn_gen_fn.
1347 * genoutput.c (output_insn_data): Cast gen_? function pointers to
1348 insn_gen_fn::stored_funcptr. Add initializer braces.
1350 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1352 Rewrite how instances of passes are cloned to remove assumptions
1353 about their sizes (thus allowing pass subclasses to have
1354 additional data fields, albeit non-GC-managed ones at this point).
1356 * passes.c (make_pass_instance): Now that passes have clone
1357 methods, rewrite this function to eliminate XNEW and memcpy
1358 calls that used hardcoded sizes. Since this function no longer
1359 creates pass instances, rename it to...
1360 (add_pass_instance): ...this. Document the old way that passes were
1361 numbered and flagged, and rework this function to continue using it.
1362 (next_pass_1): Add an initial_pass argument for use by
1364 (position_pass): When adding multiple instances of a pass, use
1365 the pass's clone method, rather than relying on the XNEW/memcpy
1366 within the former make_pass_instance (now add_pass_instance).
1367 (pass_manager::pass_manager): When invoking next_pass_1, also supply
1368 the initial instance of the current pass within the pass manager.
1370 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1372 This is the automated part of the conversion of passes from C
1373 structs to C++ classes.
1375 Patch autogenerated by refactor_passes.py from
1376 https://github.com/davidmalcolm/gcc-refactoring-scripts
1377 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
1379 * asan.c (pass_asan): Convert from a global struct to a subclass of
1380 gimple_opt_pass along with...
1381 (pass_data_asan): ...new pass_data instance and...
1382 (make_pass_asan): ...new function.
1383 (pass_asan_O0): Convert from a global struct to a subclass of
1384 gimple_opt_pass along with...
1385 (pass_data_asan_O0): ...new pass_data instance and...
1386 (make_pass_asan_O0): ...new function.
1387 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
1388 subclass of rtl_opt_pass along with...
1389 (pass_data_inc_dec): ...new pass_data instance and...
1390 (make_pass_inc_dec): ...new function.
1391 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
1392 a subclass of rtl_opt_pass along with...
1393 (pass_data_reorder_blocks): ...new pass_data instance and...
1394 (make_pass_reorder_blocks): ...new function.
1395 (pass_duplicate_computed_gotos): Convert from a global struct to a
1396 subclass of rtl_opt_pass along with...
1397 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
1398 (make_pass_duplicate_computed_gotos): ...new function.
1399 (pass_partition_blocks): Convert from a global struct to a subclass of
1400 rtl_opt_pass along with...
1401 (pass_data_partition_blocks): ...new pass_data instance and...
1402 (make_pass_partition_blocks): ...new function.
1403 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
1404 struct to a subclass of rtl_opt_pass along with...
1405 (pass_data_branch_target_load_optimize1): ...new pass_data instance
1407 (make_pass_branch_target_load_optimize1): ...new function.
1408 (pass_branch_target_load_optimize2): Convert from a global struct to a
1409 subclass of rtl_opt_pass along with...
1410 (pass_data_branch_target_load_optimize2): ...new pass_data instance
1412 (make_pass_branch_target_load_optimize2): ...new function.
1413 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
1414 of rtl_opt_pass along with...
1415 (pass_data_jump): ...new pass_data instance and...
1416 (make_pass_jump): ...new function.
1417 (pass_jump2): Convert from a global struct to a subclass of
1418 rtl_opt_pass along with...
1419 (pass_data_jump2): ...new pass_data instance and...
1420 (make_pass_jump2): ...new function.
1421 * cfgexpand.c (pass_expand): Convert from a global struct to a
1422 subclass of rtl_opt_pass along with...
1423 (pass_data_expand): ...new pass_data instance and...
1424 (make_pass_expand): ...new function.
1425 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
1426 of rtl_opt_pass along with...
1427 (pass_data_free_cfg): ...new pass_data instance and...
1428 (make_pass_free_cfg): ...new function.
1429 (pass_into_cfg_layout_mode): Convert from a global struct to a
1430 subclass of rtl_opt_pass along with...
1431 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
1432 (make_pass_into_cfg_layout_mode): ...new function.
1433 (pass_outof_cfg_layout_mode): Convert from a global struct to a
1434 subclass of rtl_opt_pass along with...
1435 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
1436 (make_pass_outof_cfg_layout_mode): ...new function.
1437 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
1438 struct to a subclass of gimple_opt_pass along with...
1439 (pass_data_build_cgraph_edges): ...new pass_data instance and...
1440 (make_pass_build_cgraph_edges): ...new function.
1441 (pass_rebuild_cgraph_edges): Convert from a global struct to a
1442 subclass of gimple_opt_pass along with...
1443 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
1444 (make_pass_rebuild_cgraph_edges): ...new function.
1445 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
1446 subclass of gimple_opt_pass along with...
1447 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
1449 (make_pass_remove_cgraph_callee_edges): ...new function.
1450 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
1451 struct to a subclass of rtl_opt_pass along with...
1452 (pass_data_stack_adjustments): ...new pass_data instance and...
1453 (make_pass_stack_adjustments): ...new function.
1454 * combine.c (pass_combine): Convert from a global struct to a subclass
1455 of rtl_opt_pass along with...
1456 (pass_data_combine): ...new pass_data instance and...
1457 (make_pass_combine): ...new function.
1458 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
1459 global struct to a subclass of rtl_opt_pass along with...
1460 (pass_data_compare_elim_after_reload): ...new pass_data instance
1462 (make_pass_compare_elim_after_reload): ...new function.
1463 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1464 of rtl_opt_pass along with...
1465 (pass_data_rtl_cprop): ...new pass_data instance and...
1466 (make_pass_rtl_cprop): ...new function.
1467 * cse.c (pass_cse): Convert from a global struct to a subclass of
1468 rtl_opt_pass along with...
1469 (pass_data_cse): ...new pass_data instance and...
1470 (make_pass_cse): ...new function.
1471 (pass_cse2): Convert from a global struct to a subclass of
1472 rtl_opt_pass along with...
1473 (pass_data_cse2): ...new pass_data instance and...
1474 (make_pass_cse2): ...new function.
1475 (pass_cse_after_global_opts): Convert from a global struct to a
1476 subclass of rtl_opt_pass along with...
1477 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1478 (make_pass_cse_after_global_opts): ...new function.
1479 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1480 of rtl_opt_pass along with...
1481 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1482 (make_pass_ud_rtl_dce): ...new function.
1483 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1484 rtl_opt_pass along with...
1485 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1486 (make_pass_fast_rtl_dce): ...new function.
1487 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1488 a subclass of rtl_opt_pass along with...
1489 (pass_data_df_initialize_opt): ...new pass_data instance and...
1490 (make_pass_df_initialize_opt): ...new function.
1491 (pass_df_initialize_no_opt): Convert from a global struct to a
1492 subclass of rtl_opt_pass along with...
1493 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1494 (make_pass_df_initialize_no_opt): ...new function.
1495 (pass_df_finish): Convert from a global struct to a subclass of
1496 rtl_opt_pass along with...
1497 (pass_data_df_finish): ...new pass_data instance and...
1498 (make_pass_df_finish): ...new function.
1499 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1500 rtl_opt_pass along with...
1501 (pass_data_rtl_dse1): ...new pass_data instance and...
1502 (make_pass_rtl_dse1): ...new function.
1503 (pass_rtl_dse2): Convert from a global struct to a subclass of
1504 rtl_opt_pass along with...
1505 (pass_data_rtl_dse2): ...new pass_data instance and...
1506 (make_pass_rtl_dse2): ...new function.
1507 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1508 subclass of rtl_opt_pass along with...
1509 (pass_data_dwarf2_frame): ...new pass_data instance and...
1510 (make_pass_dwarf2_frame): ...new function.
1511 * except.c (pass_set_nothrow_function_flags): Convert from a global
1512 struct to a subclass of rtl_opt_pass along with...
1513 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1515 (make_pass_set_nothrow_function_flags): ...new function.
1516 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1517 subclass of rtl_opt_pass along with...
1518 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1520 (make_pass_convert_to_eh_region_ranges): ...new function.
1521 * final.c (pass_compute_alignments): Convert from a global struct to a
1522 subclass of rtl_opt_pass along with...
1523 (pass_data_compute_alignments): ...new pass_data instance and...
1524 (make_pass_compute_alignments): ...new function.
1525 (pass_final): Convert from a global struct to a subclass of
1526 rtl_opt_pass along with...
1527 (pass_data_final): ...new pass_data instance and...
1528 (make_pass_final): ...new function.
1529 (pass_shorten_branches): Convert from a global struct to a subclass of
1530 rtl_opt_pass along with...
1531 (pass_data_shorten_branches): ...new pass_data instance and...
1532 (make_pass_shorten_branches): ...new function.
1533 (pass_clean_state): Convert from a global struct to a subclass of
1534 rtl_opt_pass along with...
1535 (pass_data_clean_state): ...new pass_data instance and...
1536 (make_pass_clean_state): ...new function.
1537 * function.c (pass_instantiate_virtual_regs): Convert from a global
1538 struct to a subclass of rtl_opt_pass along with...
1539 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1540 (make_pass_instantiate_virtual_regs): ...new function.
1541 (pass_leaf_regs): Convert from a global struct to a subclass of
1542 rtl_opt_pass along with...
1543 (pass_data_leaf_regs): ...new pass_data instance and...
1544 (make_pass_leaf_regs): ...new function.
1545 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1546 subclass of rtl_opt_pass along with...
1547 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1549 (make_pass_thread_prologue_and_epilogue): ...new function.
1550 (pass_match_asm_constraints): Convert from a global struct to a
1551 subclass of rtl_opt_pass along with...
1552 (pass_data_match_asm_constraints): ...new pass_data instance and...
1553 (make_pass_match_asm_constraints): ...new function.
1554 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1555 subclass of rtl_opt_pass along with...
1556 (pass_data_rtl_fwprop): ...new pass_data instance and...
1557 (make_pass_rtl_fwprop): ...new function.
1558 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1559 rtl_opt_pass along with...
1560 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1561 (make_pass_rtl_fwprop_addr): ...new function.
1562 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1563 rtl_opt_pass along with...
1564 (pass_data_rtl_pre): ...new pass_data instance and...
1565 (make_pass_rtl_pre): ...new function.
1566 (pass_rtl_hoist): Convert from a global struct to a subclass of
1567 rtl_opt_pass along with...
1568 (pass_data_rtl_hoist): ...new pass_data instance and...
1569 (make_pass_rtl_hoist): ...new function.
1570 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1571 subclass of gimple_opt_pass along with...
1572 (pass_data_lower_cf): ...new pass_data instance and...
1573 (make_pass_lower_cf): ...new function.
1574 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1575 from a global struct to a subclass of gimple_opt_pass along with...
1576 (pass_data_strength_reduction): ...new pass_data instance and...
1577 (make_pass_strength_reduction): ...new function.
1578 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1579 of rtl_opt_pass along with...
1580 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1581 (make_pass_rtl_ifcvt): ...new function.
1582 (pass_if_after_combine): Convert from a global struct to a subclass of
1583 rtl_opt_pass along with...
1584 (pass_data_if_after_combine): ...new pass_data instance and...
1585 (make_pass_if_after_combine): ...new function.
1586 (pass_if_after_reload): Convert from a global struct to a subclass of
1587 rtl_opt_pass along with...
1588 (pass_data_if_after_reload): ...new pass_data instance and...
1589 (make_pass_if_after_reload): ...new function.
1590 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1591 a subclass of rtl_opt_pass along with...
1592 (pass_data_initialize_regs): ...new pass_data instance and...
1593 (make_pass_initialize_regs): ...new function.
1594 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1595 of ipa_opt_pass_d along with...
1596 (pass_data_ipa_cp): ...new pass_data instance and...
1597 (make_pass_ipa_cp): ...new function.
1598 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1599 global struct to a subclass of gimple_opt_pass along with...
1600 (pass_data_inline_parameters): ...new pass_data instance and...
1601 (make_pass_inline_parameters): ...new function.
1602 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1603 subclass of gimple_opt_pass along with...
1604 (pass_data_early_inline): ...new pass_data instance and...
1605 (make_pass_early_inline): ...new function.
1606 (pass_ipa_inline): Convert from a global struct to a subclass of
1607 ipa_opt_pass_d along with...
1608 (pass_data_ipa_inline): ...new pass_data instance and...
1609 (make_pass_ipa_inline): ...new function.
1610 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1611 struct to a subclass of gimple_opt_pass along with...
1612 (pass_data_local_pure_const): ...new pass_data instance and...
1613 (make_pass_local_pure_const): ...new function.
1614 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1615 ipa_opt_pass_d along with...
1616 (pass_data_ipa_pure_const): ...new pass_data instance and...
1617 (make_pass_ipa_pure_const): ...new function.
1618 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1619 to a subclass of ipa_opt_pass_d along with...
1620 (pass_data_ipa_reference): ...new pass_data instance and...
1621 (make_pass_ipa_reference): ...new function.
1622 * ipa-split.c (pass_split_functions): Convert from a global struct to
1623 a subclass of gimple_opt_pass along with...
1624 (pass_data_split_functions): ...new pass_data instance and...
1625 (make_pass_split_functions): ...new function.
1626 (pass_feedback_split_functions): Convert from a global struct to a
1627 subclass of gimple_opt_pass along with...
1628 (pass_data_feedback_split_functions): ...new pass_data instance and...
1629 (make_pass_feedback_split_functions): ...new function.
1630 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1631 global struct to a subclass of simple_ipa_opt_pass along with...
1632 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1634 (make_pass_ipa_function_and_variable_visibility): ...new function.
1635 (pass_ipa_free_inline_summary): Convert from a global struct to a
1636 subclass of simple_ipa_opt_pass along with...
1637 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1638 (make_pass_ipa_free_inline_summary): ...new function.
1639 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1640 subclass of ipa_opt_pass_d along with...
1641 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1643 (make_pass_ipa_whole_program_visibility): ...new function.
1644 (pass_ipa_profile): Convert from a global struct to a subclass of
1645 ipa_opt_pass_d along with...
1646 (pass_data_ipa_profile): ...new pass_data instance and...
1647 (make_pass_ipa_profile): ...new function.
1648 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1649 ipa_opt_pass_d along with...
1650 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1651 (make_pass_ipa_cdtor_merge): ...new function.
1652 * ira.c (pass_ira): Convert from a global struct to a subclass of
1653 rtl_opt_pass along with...
1654 (pass_data_ira): ...new pass_data instance and...
1655 (make_pass_ira): ...new function.
1656 (pass_reload): Convert from a global struct to a subclass of
1657 rtl_opt_pass along with...
1658 (pass_data_reload): ...new pass_data instance and...
1659 (make_pass_reload): ...new function.
1660 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1661 subclass of rtl_opt_pass along with...
1662 (pass_data_cleanup_barriers): ...new pass_data instance and...
1663 (make_pass_cleanup_barriers): ...new function.
1664 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1665 of rtl_opt_pass along with...
1666 (pass_data_loop2): ...new pass_data instance and...
1667 (make_pass_loop2): ...new function.
1668 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1669 rtl_opt_pass along with...
1670 (pass_data_rtl_loop_init): ...new pass_data instance and...
1671 (make_pass_rtl_loop_init): ...new function.
1672 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1673 rtl_opt_pass along with...
1674 (pass_data_rtl_loop_done): ...new pass_data instance and...
1675 (make_pass_rtl_loop_done): ...new function.
1676 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1677 subclass of rtl_opt_pass along with...
1678 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1679 (make_pass_rtl_move_loop_invariants): ...new function.
1680 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1681 rtl_opt_pass along with...
1682 (pass_data_rtl_unswitch): ...new pass_data instance and...
1683 (make_pass_rtl_unswitch): ...new function.
1684 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1685 subclass of rtl_opt_pass along with...
1686 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1688 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1689 (pass_rtl_doloop): Convert from a global struct to a subclass of
1690 rtl_opt_pass along with...
1691 (pass_data_rtl_doloop): ...new pass_data instance and...
1692 (make_pass_rtl_doloop): ...new function.
1693 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1694 a subclass of rtl_opt_pass along with...
1695 (pass_data_lower_subreg): ...new pass_data instance and...
1696 (make_pass_lower_subreg): ...new function.
1697 (pass_lower_subreg2): Convert from a global struct to a subclass of
1698 rtl_opt_pass along with...
1699 (pass_data_lower_subreg2): ...new pass_data instance and...
1700 (make_pass_lower_subreg2): ...new function.
1701 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1702 struct to a subclass of ipa_opt_pass_d along with...
1703 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1704 (make_pass_ipa_lto_gimple_out): ...new function.
1705 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1706 of ipa_opt_pass_d along with...
1707 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1708 (make_pass_ipa_lto_finish_out): ...new function.
1709 * mode-switching.c (pass_mode_switching): Convert from a global struct
1710 to a subclass of rtl_opt_pass along with...
1711 (pass_data_mode_switching): ...new pass_data instance and...
1712 (make_pass_mode_switching): ...new function.
1713 * modulo-sched.c (pass_sms): Convert from a global struct to a
1714 subclass of rtl_opt_pass along with...
1715 (pass_data_sms): ...new pass_data instance and...
1716 (make_pass_sms): ...new function.
1717 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1718 subclass of gimple_opt_pass along with...
1719 (pass_data_expand_omp): ...new pass_data instance and...
1720 (make_pass_expand_omp): ...new function.
1721 (pass_lower_omp): Convert from a global struct to a subclass of
1722 gimple_opt_pass along with...
1723 (pass_data_lower_omp): ...new pass_data instance and...
1724 (make_pass_lower_omp): ...new function.
1725 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1726 of gimple_opt_pass along with...
1727 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1728 (make_pass_diagnose_omp_blocks): ...new function.
1729 * passes.c (pass_early_local_passes): Convert from a global struct to
1730 a subclass of simple_ipa_opt_pass along with...
1731 (pass_data_early_local_passes): ...new pass_data instance and...
1732 (make_pass_early_local_passes): ...new function.
1733 (pass_all_early_optimizations): Convert from a global struct to a
1734 subclass of gimple_opt_pass along with...
1735 (pass_data_all_early_optimizations): ...new pass_data instance and...
1736 (make_pass_all_early_optimizations): ...new function.
1737 (pass_all_optimizations): Convert from a global struct to a subclass
1738 of gimple_opt_pass along with...
1739 (pass_data_all_optimizations): ...new pass_data instance and...
1740 (make_pass_all_optimizations): ...new function.
1741 (pass_all_optimizations_g): Convert from a global struct to a subclass
1742 of gimple_opt_pass along with...
1743 (pass_data_all_optimizations_g): ...new pass_data instance and...
1744 (make_pass_all_optimizations_g): ...new function.
1745 (pass_rest_of_compilation): Convert from a global struct to a subclass
1746 of rtl_opt_pass along with...
1747 (pass_data_rest_of_compilation): ...new pass_data instance and...
1748 (make_pass_rest_of_compilation): ...new function.
1749 (pass_postreload): Convert from a global struct to a subclass of
1750 rtl_opt_pass along with...
1751 (pass_data_postreload): ...new pass_data instance and...
1752 (make_pass_postreload): ...new function.
1753 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1754 subclass of rtl_opt_pass along with...
1755 (pass_data_gcse2): ...new pass_data instance and...
1756 (make_pass_gcse2): ...new function.
1757 * postreload.c (pass_postreload_cse): Convert from a global struct to
1758 a subclass of rtl_opt_pass along with...
1759 (pass_data_postreload_cse): ...new pass_data instance and...
1760 (make_pass_postreload_cse): ...new function.
1761 * predict.c (pass_profile): Convert from a global struct to a subclass
1762 of gimple_opt_pass along with...
1763 (pass_data_profile): ...new pass_data instance and...
1764 (make_pass_profile): ...new function.
1765 (pass_strip_predict_hints): Convert from a global struct to a subclass
1766 of gimple_opt_pass along with...
1767 (pass_data_strip_predict_hints): ...new pass_data instance and...
1768 (make_pass_strip_predict_hints): ...new function.
1769 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1770 of rtl_opt_pass along with...
1771 (pass_data_peephole2): ...new pass_data instance and...
1772 (make_pass_peephole2): ...new function.
1773 (pass_split_all_insns): Convert from a global struct to a subclass of
1774 rtl_opt_pass along with...
1775 (pass_data_split_all_insns): ...new pass_data instance and...
1776 (make_pass_split_all_insns): ...new function.
1777 (pass_split_after_reload): Convert from a global struct to a subclass
1778 of rtl_opt_pass along with...
1779 (pass_data_split_after_reload): ...new pass_data instance and...
1780 (make_pass_split_after_reload): ...new function.
1781 (pass_split_before_regstack): Convert from a global struct to a
1782 subclass of rtl_opt_pass along with...
1783 (pass_data_split_before_regstack): ...new pass_data instance and...
1784 (make_pass_split_before_regstack): ...new function.
1785 (pass_split_before_sched2): Convert from a global struct to a subclass
1786 of rtl_opt_pass along with...
1787 (pass_data_split_before_sched2): ...new pass_data instance and...
1788 (make_pass_split_before_sched2): ...new function.
1789 (pass_split_for_shorten_branches): Convert from a global struct to a
1790 subclass of rtl_opt_pass along with...
1791 (pass_data_split_for_shorten_branches): ...new pass_data instance
1793 (make_pass_split_for_shorten_branches): ...new function.
1794 * ree.c (pass_ree): Convert from a global struct to a subclass of
1795 rtl_opt_pass along with...
1796 (pass_data_ree): ...new pass_data instance and...
1797 (make_pass_ree): ...new function.
1798 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1799 subclass of rtl_opt_pass along with...
1800 (pass_data_stack_regs): ...new pass_data instance and...
1801 (make_pass_stack_regs): ...new function.
1802 (pass_stack_regs_run): Convert from a global struct to a subclass of
1803 rtl_opt_pass along with...
1804 (pass_data_stack_regs_run): ...new pass_data instance and...
1805 (make_pass_stack_regs_run): ...new function.
1806 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1807 subclass of rtl_opt_pass along with...
1808 (pass_data_cprop_hardreg): ...new pass_data instance and...
1809 (make_pass_cprop_hardreg): ...new function.
1810 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1811 subclass of rtl_opt_pass along with...
1812 (pass_data_reginfo_init): ...new pass_data instance and...
1813 (make_pass_reginfo_init): ...new function.
1814 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1815 of rtl_opt_pass along with...
1816 (pass_data_regmove): ...new pass_data instance and...
1817 (make_pass_regmove): ...new function.
1818 * regrename.c (pass_regrename): Convert from a global struct to a
1819 subclass of rtl_opt_pass along with...
1820 (pass_data_regrename): ...new pass_data instance and...
1821 (make_pass_regrename): ...new function.
1822 * reorg.c (pass_delay_slots): Convert from a global struct to a
1823 subclass of rtl_opt_pass along with...
1824 (pass_data_delay_slots): ...new pass_data instance and...
1825 (make_pass_delay_slots): ...new function.
1826 (pass_machine_reorg): Convert from a global struct to a subclass of
1827 rtl_opt_pass along with...
1828 (pass_data_machine_reorg): ...new pass_data instance and...
1829 (make_pass_machine_reorg): ...new function.
1830 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1831 of rtl_opt_pass along with...
1832 (pass_data_sched): ...new pass_data instance and...
1833 (make_pass_sched): ...new function.
1834 (pass_sched2): Convert from a global struct to a subclass of
1835 rtl_opt_pass along with...
1836 (pass_data_sched2): ...new pass_data instance and...
1837 (make_pass_sched2): ...new function.
1838 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1839 to a subclass of rtl_opt_pass along with...
1840 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1841 (make_pass_stack_ptr_mod): ...new function.
1842 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1843 to a subclass of rtl_opt_pass along with...
1844 (pass_data_rtl_store_motion): ...new pass_data instance and...
1845 (make_pass_rtl_store_motion): ...new function.
1846 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1847 of gimple_opt_pass along with...
1848 (pass_data_tracer): ...new pass_data instance and...
1849 (make_pass_tracer): ...new function.
1850 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1851 to a subclass of gimple_opt_pass along with...
1852 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1853 (make_pass_diagnose_tm_blocks): ...new function.
1854 (pass_lower_tm): Convert from a global struct to a subclass of
1855 gimple_opt_pass along with...
1856 (pass_data_lower_tm): ...new pass_data instance and...
1857 (make_pass_lower_tm): ...new function.
1858 (pass_tm_init): Convert from a global struct to a subclass of
1859 gimple_opt_pass along with...
1860 (pass_data_tm_init): ...new pass_data instance and...
1861 (make_pass_tm_init): ...new function.
1862 (pass_tm_mark): Convert from a global struct to a subclass of
1863 gimple_opt_pass along with...
1864 (pass_data_tm_mark): ...new pass_data instance and...
1865 (make_pass_tm_mark): ...new function.
1866 (pass_tm_edges): Convert from a global struct to a subclass of
1867 gimple_opt_pass along with...
1868 (pass_data_tm_edges): ...new pass_data instance and...
1869 (make_pass_tm_edges): ...new function.
1870 (pass_tm_memopt): Convert from a global struct to a subclass of
1871 gimple_opt_pass along with...
1872 (pass_data_tm_memopt): ...new pass_data instance and...
1873 (make_pass_tm_memopt): ...new function.
1874 (pass_ipa_tm): Convert from a global struct to a subclass of
1875 simple_ipa_opt_pass along with...
1876 (pass_data_ipa_tm): ...new pass_data instance and...
1877 (make_pass_ipa_tm): ...new function.
1878 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1879 subclass of gimple_opt_pass along with...
1880 (pass_data_call_cdce): ...new pass_data instance and...
1881 (make_pass_call_cdce): ...new function.
1882 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1883 subclass of gimple_opt_pass along with...
1884 (pass_data_build_cfg): ...new pass_data instance and...
1885 (make_pass_build_cfg): ...new function.
1886 (pass_split_crit_edges): Convert from a global struct to a subclass of
1887 gimple_opt_pass along with...
1888 (pass_data_split_crit_edges): ...new pass_data instance and...
1889 (make_pass_split_crit_edges): ...new function.
1890 (pass_warn_function_return): Convert from a global struct to a
1891 subclass of gimple_opt_pass along with...
1892 (pass_data_warn_function_return): ...new pass_data instance and...
1893 (make_pass_warn_function_return): ...new function.
1894 (pass_warn_function_noreturn): Convert from a global struct to a
1895 subclass of gimple_opt_pass along with...
1896 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1897 (make_pass_warn_function_noreturn): ...new function.
1898 (pass_warn_unused_result): Convert from a global struct to a subclass
1899 of gimple_opt_pass along with...
1900 (pass_data_warn_unused_result): ...new pass_data instance and...
1901 (make_pass_warn_unused_result): ...new function.
1902 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1903 a subclass of gimple_opt_pass along with...
1904 (pass_data_merge_phi): ...new pass_data instance and...
1905 (make_pass_merge_phi): ...new function.
1906 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1907 a subclass of gimple_opt_pass along with...
1908 (pass_data_lower_complex): ...new pass_data instance and...
1909 (make_pass_lower_complex): ...new function.
1910 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1911 gimple_opt_pass along with...
1912 (pass_data_lower_complex_O0): ...new pass_data instance and...
1913 (make_pass_lower_complex_O0): ...new function.
1914 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1915 subclass of gimple_opt_pass along with...
1916 (pass_data_lower_eh): ...new pass_data instance and...
1917 (make_pass_lower_eh): ...new function.
1918 (pass_refactor_eh): Convert from a global struct to a subclass of
1919 gimple_opt_pass along with...
1920 (pass_data_refactor_eh): ...new pass_data instance and...
1921 (make_pass_refactor_eh): ...new function.
1922 (pass_lower_resx): Convert from a global struct to a subclass of
1923 gimple_opt_pass along with...
1924 (pass_data_lower_resx): ...new pass_data instance and...
1925 (make_pass_lower_resx): ...new function.
1926 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1927 of gimple_opt_pass along with...
1928 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1929 (make_pass_lower_eh_dispatch): ...new function.
1930 (pass_cleanup_eh): Convert from a global struct to a subclass of
1931 gimple_opt_pass along with...
1932 (pass_data_cleanup_eh): ...new pass_data instance and...
1933 (make_pass_cleanup_eh): ...new function.
1934 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1935 to a subclass of simple_ipa_opt_pass along with...
1936 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1937 (make_pass_ipa_lower_emutls): ...new function.
1938 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1939 a subclass of gimple_opt_pass along with...
1940 (pass_data_if_conversion): ...new pass_data instance and...
1941 (make_pass_if_conversion): ...new function.
1942 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1943 subclass of gimple_opt_pass along with...
1944 (pass_data_build_ssa): ...new pass_data instance and...
1945 (make_pass_build_ssa): ...new function.
1946 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1947 global struct to a subclass of gimple_opt_pass along with...
1948 (pass_data_loop_distribution): ...new pass_data instance and...
1949 (make_pass_loop_distribution): ...new function.
1950 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1951 subclass of gimple_opt_pass along with...
1952 (pass_data_mudflap_1): ...new pass_data instance and...
1953 (make_pass_mudflap_1): ...new function.
1954 (pass_mudflap_2): Convert from a global struct to a subclass of
1955 gimple_opt_pass along with...
1956 (pass_data_mudflap_2): ...new pass_data instance and...
1957 (make_pass_mudflap_2): ...new function.
1958 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1959 subclass of gimple_opt_pass along with...
1960 (pass_data_mudflap_1): ...new pass_data instance and...
1961 (make_pass_mudflap_1): ...new function.
1962 (pass_mudflap_2): Convert from a global struct to a subclass of
1963 gimple_opt_pass along with...
1964 (pass_data_mudflap_2): ...new pass_data instance and...
1965 (make_pass_mudflap_2): ...new function.
1966 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1967 gimple_opt_pass along with...
1968 (pass_data_nrv): ...new pass_data instance and...
1969 (make_pass_nrv): ...new function.
1970 (pass_return_slot): Convert from a global struct to a subclass of
1971 gimple_opt_pass along with...
1972 (pass_data_return_slot): ...new pass_data instance and...
1973 (make_pass_return_slot): ...new function.
1974 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1975 to a subclass of gimple_opt_pass along with...
1976 (pass_data_object_sizes): ...new pass_data instance and...
1977 (make_pass_object_sizes): ...new function.
1978 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1979 global struct to a subclass of gimple_opt_pass along with...
1980 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1982 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1983 (pass_fixup_cfg): Convert from a global struct to a subclass of
1984 gimple_opt_pass along with...
1985 (pass_data_fixup_cfg): ...new pass_data instance and...
1986 (make_pass_fixup_cfg): ...new function.
1987 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1988 (make_pass_mudflap_1): ...new function.
1989 (pass_mudflap_2): Replace declaration with that of...
1990 (make_pass_mudflap_2): ...new function.
1991 (pass_asan): Replace declaration with that of...
1992 (make_pass_asan): ...new function.
1993 (pass_asan_O0): Replace declaration with that of...
1994 (make_pass_asan_O0): ...new function.
1995 (pass_tsan): Replace declaration with that of...
1996 (make_pass_tsan): ...new function.
1997 (pass_tsan_O0): Replace declaration with that of...
1998 (make_pass_tsan_O0): ...new function.
1999 (pass_lower_cf): Replace declaration with that of...
2000 (make_pass_lower_cf): ...new function.
2001 (pass_refactor_eh): Replace declaration with that of...
2002 (make_pass_refactor_eh): ...new function.
2003 (pass_lower_eh): Replace declaration with that of...
2004 (make_pass_lower_eh): ...new function.
2005 (pass_lower_eh_dispatch): Replace declaration with that of...
2006 (make_pass_lower_eh_dispatch): ...new function.
2007 (pass_lower_resx): Replace declaration with that of...
2008 (make_pass_lower_resx): ...new function.
2009 (pass_build_cfg): Replace declaration with that of...
2010 (make_pass_build_cfg): ...new function.
2011 (pass_early_tree_profile): Replace declaration with that of...
2012 (make_pass_early_tree_profile): ...new function.
2013 (pass_cleanup_eh): Replace declaration with that of...
2014 (make_pass_cleanup_eh): ...new function.
2015 (pass_sra): Replace declaration with that of...
2016 (make_pass_sra): ...new function.
2017 (pass_sra_early): Replace declaration with that of...
2018 (make_pass_sra_early): ...new function.
2019 (pass_early_ipa_sra): Replace declaration with that of...
2020 (make_pass_early_ipa_sra): ...new function.
2021 (pass_tail_recursion): Replace declaration with that of...
2022 (make_pass_tail_recursion): ...new function.
2023 (pass_tail_calls): Replace declaration with that of...
2024 (make_pass_tail_calls): ...new function.
2025 (pass_tree_loop): Replace declaration with that of...
2026 (make_pass_tree_loop): ...new function.
2027 (pass_tree_loop_init): Replace declaration with that of...
2028 (make_pass_tree_loop_init): ...new function.
2029 (pass_lim): Replace declaration with that of...
2030 (make_pass_lim): ...new function.
2031 (pass_tree_unswitch): Replace declaration with that of...
2032 (make_pass_tree_unswitch): ...new function.
2033 (pass_predcom): Replace declaration with that of...
2034 (make_pass_predcom): ...new function.
2035 (pass_iv_canon): Replace declaration with that of...
2036 (make_pass_iv_canon): ...new function.
2037 (pass_scev_cprop): Replace declaration with that of...
2038 (make_pass_scev_cprop): ...new function.
2039 (pass_empty_loop): Replace declaration with that of...
2040 (make_pass_empty_loop): ...new function.
2041 (pass_record_bounds): Replace declaration with that of...
2042 (make_pass_record_bounds): ...new function.
2043 (pass_graphite): Replace declaration with that of...
2044 (make_pass_graphite): ...new function.
2045 (pass_graphite_transforms): Replace declaration with that of...
2046 (make_pass_graphite_transforms): ...new function.
2047 (pass_if_conversion): Replace declaration with that of...
2048 (make_pass_if_conversion): ...new function.
2049 (pass_loop_distribution): Replace declaration with that of...
2050 (make_pass_loop_distribution): ...new function.
2051 (pass_vectorize): Replace declaration with that of...
2052 (make_pass_vectorize): ...new function.
2053 (pass_slp_vectorize): Replace declaration with that of...
2054 (make_pass_slp_vectorize): ...new function.
2055 (pass_complete_unroll): Replace declaration with that of...
2056 (make_pass_complete_unroll): ...new function.
2057 (pass_complete_unrolli): Replace declaration with that of...
2058 (make_pass_complete_unrolli): ...new function.
2059 (pass_parallelize_loops): Replace declaration with that of...
2060 (make_pass_parallelize_loops): ...new function.
2061 (pass_loop_prefetch): Replace declaration with that of...
2062 (make_pass_loop_prefetch): ...new function.
2063 (pass_iv_optimize): Replace declaration with that of...
2064 (make_pass_iv_optimize): ...new function.
2065 (pass_tree_loop_done): Replace declaration with that of...
2066 (make_pass_tree_loop_done): ...new function.
2067 (pass_ch): Replace declaration with that of...
2068 (make_pass_ch): ...new function.
2069 (pass_ccp): Replace declaration with that of...
2070 (make_pass_ccp): ...new function.
2071 (pass_phi_only_cprop): Replace declaration with that of...
2072 (make_pass_phi_only_cprop): ...new function.
2073 (pass_build_ssa): Replace declaration with that of...
2074 (make_pass_build_ssa): ...new function.
2075 (pass_build_alias): Replace declaration with that of...
2076 (make_pass_build_alias): ...new function.
2077 (pass_build_ealias): Replace declaration with that of...
2078 (make_pass_build_ealias): ...new function.
2079 (pass_dominator): Replace declaration with that of...
2080 (make_pass_dominator): ...new function.
2081 (pass_dce): Replace declaration with that of...
2082 (make_pass_dce): ...new function.
2083 (pass_dce_loop): Replace declaration with that of...
2084 (make_pass_dce_loop): ...new function.
2085 (pass_cd_dce): Replace declaration with that of...
2086 (make_pass_cd_dce): ...new function.
2087 (pass_call_cdce): Replace declaration with that of...
2088 (make_pass_call_cdce): ...new function.
2089 (pass_merge_phi): Replace declaration with that of...
2090 (make_pass_merge_phi): ...new function.
2091 (pass_split_crit_edges): Replace declaration with that of...
2092 (make_pass_split_crit_edges): ...new function.
2093 (pass_pre): Replace declaration with that of...
2094 (make_pass_pre): ...new function.
2095 (pass_profile): Replace declaration with that of...
2096 (make_pass_profile): ...new function.
2097 (pass_strip_predict_hints): Replace declaration with that of...
2098 (make_pass_strip_predict_hints): ...new function.
2099 (pass_lower_complex_O0): Replace declaration with that of...
2100 (make_pass_lower_complex_O0): ...new function.
2101 (pass_lower_complex): Replace declaration with that of...
2102 (make_pass_lower_complex): ...new function.
2103 (pass_lower_vector): Replace declaration with that of...
2104 (make_pass_lower_vector): ...new function.
2105 (pass_lower_vector_ssa): Replace declaration with that of...
2106 (make_pass_lower_vector_ssa): ...new function.
2107 (pass_lower_omp): Replace declaration with that of...
2108 (make_pass_lower_omp): ...new function.
2109 (pass_diagnose_omp_blocks): Replace declaration with that of...
2110 (make_pass_diagnose_omp_blocks): ...new function.
2111 (pass_expand_omp): Replace declaration with that of...
2112 (make_pass_expand_omp): ...new function.
2113 (pass_expand_omp_ssa): Replace declaration with that of...
2114 (make_pass_expand_omp_ssa): ...new function.
2115 (pass_object_sizes): Replace declaration with that of...
2116 (make_pass_object_sizes): ...new function.
2117 (pass_strlen): Replace declaration with that of...
2118 (make_pass_strlen): ...new function.
2119 (pass_fold_builtins): Replace declaration with that of...
2120 (make_pass_fold_builtins): ...new function.
2121 (pass_stdarg): Replace declaration with that of...
2122 (make_pass_stdarg): ...new function.
2123 (pass_early_warn_uninitialized): Replace declaration with that of...
2124 (make_pass_early_warn_uninitialized): ...new function.
2125 (pass_late_warn_uninitialized): Replace declaration with that of...
2126 (make_pass_late_warn_uninitialized): ...new function.
2127 (pass_cse_reciprocals): Replace declaration with that of...
2128 (make_pass_cse_reciprocals): ...new function.
2129 (pass_cse_sincos): Replace declaration with that of...
2130 (make_pass_cse_sincos): ...new function.
2131 (pass_optimize_bswap): Replace declaration with that of...
2132 (make_pass_optimize_bswap): ...new function.
2133 (pass_optimize_widening_mul): Replace declaration with that of...
2134 (make_pass_optimize_widening_mul): ...new function.
2135 (pass_warn_function_return): Replace declaration with that of...
2136 (make_pass_warn_function_return): ...new function.
2137 (pass_warn_function_noreturn): Replace declaration with that of...
2138 (make_pass_warn_function_noreturn): ...new function.
2139 (pass_cselim): Replace declaration with that of...
2140 (make_pass_cselim): ...new function.
2141 (pass_phiopt): Replace declaration with that of...
2142 (make_pass_phiopt): ...new function.
2143 (pass_forwprop): Replace declaration with that of...
2144 (make_pass_forwprop): ...new function.
2145 (pass_phiprop): Replace declaration with that of...
2146 (make_pass_phiprop): ...new function.
2147 (pass_tree_ifcombine): Replace declaration with that of...
2148 (make_pass_tree_ifcombine): ...new function.
2149 (pass_dse): Replace declaration with that of...
2150 (make_pass_dse): ...new function.
2151 (pass_nrv): Replace declaration with that of...
2152 (make_pass_nrv): ...new function.
2153 (pass_rename_ssa_copies): Replace declaration with that of...
2154 (make_pass_rename_ssa_copies): ...new function.
2155 (pass_sink_code): Replace declaration with that of...
2156 (make_pass_sink_code): ...new function.
2157 (pass_fre): Replace declaration with that of...
2158 (make_pass_fre): ...new function.
2159 (pass_check_data_deps): Replace declaration with that of...
2160 (make_pass_check_data_deps): ...new function.
2161 (pass_copy_prop): Replace declaration with that of...
2162 (make_pass_copy_prop): ...new function.
2163 (pass_vrp): Replace declaration with that of...
2164 (make_pass_vrp): ...new function.
2165 (pass_uncprop): Replace declaration with that of...
2166 (make_pass_uncprop): ...new function.
2167 (pass_return_slot): Replace declaration with that of...
2168 (make_pass_return_slot): ...new function.
2169 (pass_reassoc): Replace declaration with that of...
2170 (make_pass_reassoc): ...new function.
2171 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2172 (make_pass_rebuild_cgraph_edges): ...new function.
2173 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2174 (make_pass_remove_cgraph_callee_edges): ...new function.
2175 (pass_build_cgraph_edges): Replace declaration with that of...
2176 (make_pass_build_cgraph_edges): ...new function.
2177 (pass_local_pure_const): Replace declaration with that of...
2178 (make_pass_local_pure_const): ...new function.
2179 (pass_tracer): Replace declaration with that of...
2180 (make_pass_tracer): ...new function.
2181 (pass_warn_unused_result): Replace declaration with that of...
2182 (make_pass_warn_unused_result): ...new function.
2183 (pass_diagnose_tm_blocks): Replace declaration with that of...
2184 (make_pass_diagnose_tm_blocks): ...new function.
2185 (pass_lower_tm): Replace declaration with that of...
2186 (make_pass_lower_tm): ...new function.
2187 (pass_tm_init): Replace declaration with that of...
2188 (make_pass_tm_init): ...new function.
2189 (pass_tm_mark): Replace declaration with that of...
2190 (make_pass_tm_mark): ...new function.
2191 (pass_tm_memopt): Replace declaration with that of...
2192 (make_pass_tm_memopt): ...new function.
2193 (pass_tm_edges): Replace declaration with that of...
2194 (make_pass_tm_edges): ...new function.
2195 (pass_split_functions): Replace declaration with that of...
2196 (make_pass_split_functions): ...new function.
2197 (pass_feedback_split_functions): Replace declaration with that of...
2198 (make_pass_feedback_split_functions): ...new function.
2199 (pass_strength_reduction): Replace declaration with that of...
2200 (make_pass_strength_reduction): ...new function.
2201 (pass_ipa_lower_emutls): Replace declaration with that of...
2202 (make_pass_ipa_lower_emutls): ...new function.
2203 (pass_ipa_function_and_variable_visibility): Replace declaration with
2205 (make_pass_ipa_function_and_variable_visibility): ...new function.
2206 (pass_ipa_tree_profile): Replace declaration with that of...
2207 (make_pass_ipa_tree_profile): ...new function.
2208 (pass_early_local_passes): Replace declaration with that of...
2209 (make_pass_early_local_passes): ...new function.
2210 (pass_ipa_whole_program_visibility): Replace declaration with that
2212 (make_pass_ipa_whole_program_visibility): ...new function.
2213 (pass_ipa_lto_gimple_out): Replace declaration with that of...
2214 (make_pass_ipa_lto_gimple_out): ...new function.
2215 (pass_ipa_increase_alignment): Replace declaration with that of...
2216 (make_pass_ipa_increase_alignment): ...new function.
2217 (pass_ipa_inline): Replace declaration with that of...
2218 (make_pass_ipa_inline): ...new function.
2219 (pass_ipa_free_lang_data): Replace declaration with that of...
2220 (make_pass_ipa_free_lang_data): ...new function.
2221 (pass_ipa_free_inline_summary): Replace declaration with that of...
2222 (make_pass_ipa_free_inline_summary): ...new function.
2223 (pass_ipa_cp): Replace declaration with that of...
2224 (make_pass_ipa_cp): ...new function.
2225 (pass_ipa_reference): Replace declaration with that of...
2226 (make_pass_ipa_reference): ...new function.
2227 (pass_ipa_pure_const): Replace declaration with that of...
2228 (make_pass_ipa_pure_const): ...new function.
2229 (pass_ipa_pta): Replace declaration with that of...
2230 (make_pass_ipa_pta): ...new function.
2231 (pass_ipa_lto_finish_out): Replace declaration with that of...
2232 (make_pass_ipa_lto_finish_out): ...new function.
2233 (pass_ipa_tm): Replace declaration with that of...
2234 (make_pass_ipa_tm): ...new function.
2235 (pass_ipa_profile): Replace declaration with that of...
2236 (make_pass_ipa_profile): ...new function.
2237 (pass_ipa_cdtor_merge): Replace declaration with that of...
2238 (make_pass_ipa_cdtor_merge): ...new function.
2239 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
2241 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2242 (pass_init_datastructures): Replace declaration with that of...
2243 (make_pass_init_datastructures): ...new function.
2244 (pass_fixup_cfg): Replace declaration with that of...
2245 (make_pass_fixup_cfg): ...new function.
2246 (pass_expand): Replace declaration with that of...
2247 (make_pass_expand): ...new function.
2248 (pass_instantiate_virtual_regs): Replace declaration with that of...
2249 (make_pass_instantiate_virtual_regs): ...new function.
2250 (pass_rtl_fwprop): Replace declaration with that of...
2251 (make_pass_rtl_fwprop): ...new function.
2252 (pass_rtl_fwprop_addr): Replace declaration with that of...
2253 (make_pass_rtl_fwprop_addr): ...new function.
2254 (pass_jump): Replace declaration with that of...
2255 (make_pass_jump): ...new function.
2256 (pass_jump2): Replace declaration with that of...
2257 (make_pass_jump2): ...new function.
2258 (pass_lower_subreg): Replace declaration with that of...
2259 (make_pass_lower_subreg): ...new function.
2260 (pass_cse): Replace declaration with that of...
2261 (make_pass_cse): ...new function.
2262 (pass_fast_rtl_dce): Replace declaration with that of...
2263 (make_pass_fast_rtl_dce): ...new function.
2264 (pass_ud_rtl_dce): Replace declaration with that of...
2265 (make_pass_ud_rtl_dce): ...new function.
2266 (pass_rtl_dce): Replace declaration with that of...
2267 (make_pass_rtl_dce): ...new function.
2268 (pass_rtl_dse1): Replace declaration with that of...
2269 (make_pass_rtl_dse1): ...new function.
2270 (pass_rtl_dse2): Replace declaration with that of...
2271 (make_pass_rtl_dse2): ...new function.
2272 (pass_rtl_dse3): Replace declaration with that of...
2273 (make_pass_rtl_dse3): ...new function.
2274 (pass_rtl_cprop): Replace declaration with that of...
2275 (make_pass_rtl_cprop): ...new function.
2276 (pass_rtl_pre): Replace declaration with that of...
2277 (make_pass_rtl_pre): ...new function.
2278 (pass_rtl_hoist): Replace declaration with that of...
2279 (make_pass_rtl_hoist): ...new function.
2280 (pass_rtl_store_motion): Replace declaration with that of...
2281 (make_pass_rtl_store_motion): ...new function.
2282 (pass_cse_after_global_opts): Replace declaration with that of...
2283 (make_pass_cse_after_global_opts): ...new function.
2284 (pass_rtl_ifcvt): Replace declaration with that of...
2285 (make_pass_rtl_ifcvt): ...new function.
2286 (pass_into_cfg_layout_mode): Replace declaration with that of...
2287 (make_pass_into_cfg_layout_mode): ...new function.
2288 (pass_outof_cfg_layout_mode): Replace declaration with that of...
2289 (make_pass_outof_cfg_layout_mode): ...new function.
2290 (pass_loop2): Replace declaration with that of...
2291 (make_pass_loop2): ...new function.
2292 (pass_rtl_loop_init): Replace declaration with that of...
2293 (make_pass_rtl_loop_init): ...new function.
2294 (pass_rtl_move_loop_invariants): Replace declaration with that of...
2295 (make_pass_rtl_move_loop_invariants): ...new function.
2296 (pass_rtl_unswitch): Replace declaration with that of...
2297 (make_pass_rtl_unswitch): ...new function.
2298 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
2299 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2300 (pass_rtl_doloop): Replace declaration with that of...
2301 (make_pass_rtl_doloop): ...new function.
2302 (pass_rtl_loop_done): Replace declaration with that of...
2303 (make_pass_rtl_loop_done): ...new function.
2304 (pass_web): Replace declaration with that of...
2305 (make_pass_web): ...new function.
2306 (pass_cse2): Replace declaration with that of...
2307 (make_pass_cse2): ...new function.
2308 (pass_df_initialize_opt): Replace declaration with that of...
2309 (make_pass_df_initialize_opt): ...new function.
2310 (pass_df_initialize_no_opt): Replace declaration with that of...
2311 (make_pass_df_initialize_no_opt): ...new function.
2312 (pass_reginfo_init): Replace declaration with that of...
2313 (make_pass_reginfo_init): ...new function.
2314 (pass_inc_dec): Replace declaration with that of...
2315 (make_pass_inc_dec): ...new function.
2316 (pass_stack_ptr_mod): Replace declaration with that of...
2317 (make_pass_stack_ptr_mod): ...new function.
2318 (pass_initialize_regs): Replace declaration with that of...
2319 (make_pass_initialize_regs): ...new function.
2320 (pass_combine): Replace declaration with that of...
2321 (make_pass_combine): ...new function.
2322 (pass_if_after_combine): Replace declaration with that of...
2323 (make_pass_if_after_combine): ...new function.
2324 (pass_ree): Replace declaration with that of...
2325 (make_pass_ree): ...new function.
2326 (pass_partition_blocks): Replace declaration with that of...
2327 (make_pass_partition_blocks): ...new function.
2328 (pass_match_asm_constraints): Replace declaration with that of...
2329 (make_pass_match_asm_constraints): ...new function.
2330 (pass_regmove): Replace declaration with that of...
2331 (make_pass_regmove): ...new function.
2332 (pass_split_all_insns): Replace declaration with that of...
2333 (make_pass_split_all_insns): ...new function.
2334 (pass_fast_rtl_byte_dce): Replace declaration with that of...
2335 (make_pass_fast_rtl_byte_dce): ...new function.
2336 (pass_lower_subreg2): Replace declaration with that of...
2337 (make_pass_lower_subreg2): ...new function.
2338 (pass_mode_switching): Replace declaration with that of...
2339 (make_pass_mode_switching): ...new function.
2340 (pass_sms): Replace declaration with that of...
2341 (make_pass_sms): ...new function.
2342 (pass_sched): Replace declaration with that of...
2343 (make_pass_sched): ...new function.
2344 (pass_ira): Replace declaration with that of...
2345 (make_pass_ira): ...new function.
2346 (pass_reload): Replace declaration with that of...
2347 (make_pass_reload): ...new function.
2348 (pass_clean_state): Replace declaration with that of...
2349 (make_pass_clean_state): ...new function.
2350 (pass_branch_prob): Replace declaration with that of...
2351 (make_pass_branch_prob): ...new function.
2352 (pass_value_profile_transformations): Replace declaration with that
2354 (make_pass_value_profile_transformations): ...new function.
2355 (pass_postreload_cse): Replace declaration with that of...
2356 (make_pass_postreload_cse): ...new function.
2357 (pass_gcse2): Replace declaration with that of...
2358 (make_pass_gcse2): ...new function.
2359 (pass_split_after_reload): Replace declaration with that of...
2360 (make_pass_split_after_reload): ...new function.
2361 (pass_branch_target_load_optimize1): Replace declaration with that
2363 (make_pass_branch_target_load_optimize1): ...new function.
2364 (pass_thread_prologue_and_epilogue): Replace declaration with that
2366 (make_pass_thread_prologue_and_epilogue): ...new function.
2367 (pass_stack_adjustments): Replace declaration with that of...
2368 (make_pass_stack_adjustments): ...new function.
2369 (pass_peephole2): Replace declaration with that of...
2370 (make_pass_peephole2): ...new function.
2371 (pass_if_after_reload): Replace declaration with that of...
2372 (make_pass_if_after_reload): ...new function.
2373 (pass_regrename): Replace declaration with that of...
2374 (make_pass_regrename): ...new function.
2375 (pass_cprop_hardreg): Replace declaration with that of...
2376 (make_pass_cprop_hardreg): ...new function.
2377 (pass_reorder_blocks): Replace declaration with that of...
2378 (make_pass_reorder_blocks): ...new function.
2379 (pass_branch_target_load_optimize2): Replace declaration with that
2381 (make_pass_branch_target_load_optimize2): ...new function.
2382 (pass_leaf_regs): Replace declaration with that of...
2383 (make_pass_leaf_regs): ...new function.
2384 (pass_split_before_sched2): Replace declaration with that of...
2385 (make_pass_split_before_sched2): ...new function.
2386 (pass_compare_elim_after_reload): Replace declaration with that of...
2387 (make_pass_compare_elim_after_reload): ...new function.
2388 (pass_sched2): Replace declaration with that of...
2389 (make_pass_sched2): ...new function.
2390 (pass_stack_regs): Replace declaration with that of...
2391 (make_pass_stack_regs): ...new function.
2392 (pass_stack_regs_run): Replace declaration with that of...
2393 (make_pass_stack_regs_run): ...new function.
2394 (pass_df_finish): Replace declaration with that of...
2395 (make_pass_df_finish): ...new function.
2396 (pass_compute_alignments): Replace declaration with that of...
2397 (make_pass_compute_alignments): ...new function.
2398 (pass_duplicate_computed_gotos): Replace declaration with that of...
2399 (make_pass_duplicate_computed_gotos): ...new function.
2400 (pass_variable_tracking): Replace declaration with that of...
2401 (make_pass_variable_tracking): ...new function.
2402 (pass_free_cfg): Replace declaration with that of...
2403 (make_pass_free_cfg): ...new function.
2404 (pass_machine_reorg): Replace declaration with that of...
2405 (make_pass_machine_reorg): ...new function.
2406 (pass_cleanup_barriers): Replace declaration with that of...
2407 (make_pass_cleanup_barriers): ...new function.
2408 (pass_delay_slots): Replace declaration with that of...
2409 (make_pass_delay_slots): ...new function.
2410 (pass_split_for_shorten_branches): Replace declaration with that of...
2411 (make_pass_split_for_shorten_branches): ...new function.
2412 (pass_split_before_regstack): Replace declaration with that of...
2413 (make_pass_split_before_regstack): ...new function.
2414 (pass_convert_to_eh_region_ranges): Replace declaration with that
2416 (make_pass_convert_to_eh_region_ranges): ...new function.
2417 (pass_shorten_branches): Replace declaration with that of...
2418 (make_pass_shorten_branches): ...new function.
2419 (pass_set_nothrow_function_flags): Replace declaration with that of...
2420 (make_pass_set_nothrow_function_flags): ...new function.
2421 (pass_dwarf2_frame): Replace declaration with that of...
2422 (make_pass_dwarf2_frame): ...new function.
2423 (pass_final): Replace declaration with that of...
2424 (make_pass_final): ...new function.
2425 (pass_rtl_seqabstr): Replace declaration with that of...
2426 (make_pass_rtl_seqabstr): ...new function.
2427 (pass_release_ssa_names): Replace declaration with that of...
2428 (make_pass_release_ssa_names): ...new function.
2429 (pass_early_inline): Replace declaration with that of...
2430 (make_pass_early_inline): ...new function.
2431 (pass_inline_parameters): Replace declaration with that of...
2432 (make_pass_inline_parameters): ...new function.
2433 (pass_update_address_taken): Replace declaration with that of...
2434 (make_pass_update_address_taken): ...new function.
2435 (pass_convert_switch): Replace declaration with that of...
2436 (make_pass_convert_switch): ...new function.
2437 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
2438 to a subclass of simple_ipa_opt_pass along with...
2439 (pass_data_ipa_tree_profile): ...new pass_data instance and...
2440 (make_pass_ipa_tree_profile): ...new function.
2441 * tree-sra.c (pass_sra_early): Convert from a global struct to a
2442 subclass of gimple_opt_pass along with...
2443 (pass_data_sra_early): ...new pass_data instance and...
2444 (make_pass_sra_early): ...new function.
2445 (pass_sra): Convert from a global struct to a subclass of
2446 gimple_opt_pass along with...
2447 (pass_data_sra): ...new pass_data instance and...
2448 (make_pass_sra): ...new function.
2449 (pass_early_ipa_sra): Convert from a global struct to a subclass of
2450 gimple_opt_pass along with...
2451 (pass_data_early_ipa_sra): ...new pass_data instance and...
2452 (make_pass_early_ipa_sra): ...new function.
2453 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
2454 subclass of gimple_opt_pass along with...
2455 (pass_data_ccp): ...new pass_data instance and...
2456 (make_pass_ccp): ...new function.
2457 (pass_fold_builtins): Convert from a global struct to a subclass of
2458 gimple_opt_pass along with...
2459 (pass_data_fold_builtins): ...new pass_data instance and...
2460 (make_pass_fold_builtins): ...new function.
2461 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2462 subclass of gimple_opt_pass along with...
2463 (pass_data_copy_prop): ...new pass_data instance and...
2464 (make_pass_copy_prop): ...new function.
2465 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2466 global struct to a subclass of gimple_opt_pass along with...
2467 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2468 (make_pass_rename_ssa_copies): ...new function.
2469 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2470 subclass of gimple_opt_pass along with...
2471 (pass_data_dce): ...new pass_data instance and...
2472 (make_pass_dce): ...new function.
2473 (pass_dce_loop): Convert from a global struct to a subclass of
2474 gimple_opt_pass along with...
2475 (pass_data_dce_loop): ...new pass_data instance and...
2476 (make_pass_dce_loop): ...new function.
2477 (pass_cd_dce): Convert from a global struct to a subclass of
2478 gimple_opt_pass along with...
2479 (pass_data_cd_dce): ...new pass_data instance and...
2480 (make_pass_cd_dce): ...new function.
2481 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2482 subclass of gimple_opt_pass along with...
2483 (pass_data_dominator): ...new pass_data instance and...
2484 (make_pass_dominator): ...new function.
2485 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2486 gimple_opt_pass along with...
2487 (pass_data_phi_only_cprop): ...new pass_data instance and...
2488 (make_pass_phi_only_cprop): ...new function.
2489 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2490 subclass of gimple_opt_pass along with...
2491 (pass_data_dse): ...new pass_data instance and...
2492 (make_pass_dse): ...new function.
2493 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2494 a subclass of gimple_opt_pass along with...
2495 (pass_data_forwprop): ...new pass_data instance and...
2496 (make_pass_forwprop): ...new function.
2497 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2498 struct to a subclass of gimple_opt_pass along with...
2499 (pass_data_tree_ifcombine): ...new pass_data instance and...
2500 (make_pass_tree_ifcombine): ...new function.
2501 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2502 subclass of gimple_opt_pass along with...
2503 (pass_data_ch): ...new pass_data instance and...
2504 (make_pass_ch): ...new function.
2505 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2506 subclass of gimple_opt_pass along with...
2507 (pass_data_tree_loop): ...new pass_data instance and...
2508 (make_pass_tree_loop): ...new function.
2509 (pass_tree_loop_init): Convert from a global struct to a subclass of
2510 gimple_opt_pass along with...
2511 (pass_data_tree_loop_init): ...new pass_data instance and...
2512 (make_pass_tree_loop_init): ...new function.
2513 (pass_lim): Convert from a global struct to a subclass of
2514 gimple_opt_pass along with...
2515 (pass_data_lim): ...new pass_data instance and...
2516 (make_pass_lim): ...new function.
2517 (pass_tree_unswitch): Convert from a global struct to a subclass of
2518 gimple_opt_pass along with...
2519 (pass_data_tree_unswitch): ...new pass_data instance and...
2520 (make_pass_tree_unswitch): ...new function.
2521 (pass_predcom): Convert from a global struct to a subclass of
2522 gimple_opt_pass along with...
2523 (pass_data_predcom): ...new pass_data instance and...
2524 (make_pass_predcom): ...new function.
2525 (pass_vectorize): Convert from a global struct to a subclass of
2526 gimple_opt_pass along with...
2527 (pass_data_vectorize): ...new pass_data instance and...
2528 (make_pass_vectorize): ...new function.
2529 (pass_graphite): Convert from a global struct to a subclass of
2530 gimple_opt_pass along with...
2531 (pass_data_graphite): ...new pass_data instance and...
2532 (make_pass_graphite): ...new function.
2533 (pass_graphite_transforms): Convert from a global struct to a subclass
2534 of gimple_opt_pass along with...
2535 (pass_data_graphite_transforms): ...new pass_data instance and...
2536 (make_pass_graphite_transforms): ...new function.
2537 (pass_check_data_deps): Convert from a global struct to a subclass of
2538 gimple_opt_pass along with...
2539 (pass_data_check_data_deps): ...new pass_data instance and...
2540 (make_pass_check_data_deps): ...new function.
2541 (pass_iv_canon): Convert from a global struct to a subclass of
2542 gimple_opt_pass along with...
2543 (pass_data_iv_canon): ...new pass_data instance and...
2544 (make_pass_iv_canon): ...new function.
2545 (pass_scev_cprop): Convert from a global struct to a subclass of
2546 gimple_opt_pass along with...
2547 (pass_data_scev_cprop): ...new pass_data instance and...
2548 (make_pass_scev_cprop): ...new function.
2549 (pass_record_bounds): Convert from a global struct to a subclass of
2550 gimple_opt_pass along with...
2551 (pass_data_record_bounds): ...new pass_data instance and...
2552 (make_pass_record_bounds): ...new function.
2553 (pass_complete_unroll): Convert from a global struct to a subclass of
2554 gimple_opt_pass along with...
2555 (pass_data_complete_unroll): ...new pass_data instance and...
2556 (make_pass_complete_unroll): ...new function.
2557 (pass_complete_unrolli): Convert from a global struct to a subclass of
2558 gimple_opt_pass along with...
2559 (pass_data_complete_unrolli): ...new pass_data instance and...
2560 (make_pass_complete_unrolli): ...new function.
2561 (pass_parallelize_loops): Convert from a global struct to a subclass
2562 of gimple_opt_pass along with...
2563 (pass_data_parallelize_loops): ...new pass_data instance and...
2564 (make_pass_parallelize_loops): ...new function.
2565 (pass_loop_prefetch): Convert from a global struct to a subclass of
2566 gimple_opt_pass along with...
2567 (pass_data_loop_prefetch): ...new pass_data instance and...
2568 (make_pass_loop_prefetch): ...new function.
2569 (pass_iv_optimize): Convert from a global struct to a subclass of
2570 gimple_opt_pass along with...
2571 (pass_data_iv_optimize): ...new pass_data instance and...
2572 (make_pass_iv_optimize): ...new function.
2573 (pass_tree_loop_done): Convert from a global struct to a subclass of
2574 gimple_opt_pass along with...
2575 (pass_data_tree_loop_done): ...new pass_data instance and...
2576 (make_pass_tree_loop_done): ...new function.
2577 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2578 struct to a subclass of gimple_opt_pass along with...
2579 (pass_data_cse_reciprocals): ...new pass_data instance and...
2580 (make_pass_cse_reciprocals): ...new function.
2581 (pass_cse_sincos): Convert from a global struct to a subclass of
2582 gimple_opt_pass along with...
2583 (pass_data_cse_sincos): ...new pass_data instance and...
2584 (make_pass_cse_sincos): ...new function.
2585 (pass_optimize_bswap): Convert from a global struct to a subclass of
2586 gimple_opt_pass along with...
2587 (pass_data_optimize_bswap): ...new pass_data instance and...
2588 (make_pass_optimize_bswap): ...new function.
2589 (pass_optimize_widening_mul): Convert from a global struct to a
2590 subclass of gimple_opt_pass along with...
2591 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2592 (make_pass_optimize_widening_mul): ...new function.
2593 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2594 subclass of gimple_opt_pass along with...
2595 (pass_data_phiopt): ...new pass_data instance and...
2596 (make_pass_phiopt): ...new function.
2597 (pass_cselim): Convert from a global struct to a subclass of
2598 gimple_opt_pass along with...
2599 (pass_data_cselim): ...new pass_data instance and...
2600 (make_pass_cselim): ...new function.
2601 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2602 subclass of gimple_opt_pass along with...
2603 (pass_data_phiprop): ...new pass_data instance and...
2604 (make_pass_phiprop): ...new function.
2605 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2606 subclass of gimple_opt_pass along with...
2607 (pass_data_pre): ...new pass_data instance and...
2608 (make_pass_pre): ...new function.
2609 (pass_fre): Convert from a global struct to a subclass of
2610 gimple_opt_pass along with...
2611 (pass_data_fre): ...new pass_data instance and...
2612 (make_pass_fre): ...new function.
2613 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2614 subclass of gimple_opt_pass along with...
2615 (pass_data_reassoc): ...new pass_data instance and...
2616 (make_pass_reassoc): ...new function.
2617 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2618 subclass of gimple_opt_pass along with...
2619 (pass_data_sink_code): ...new pass_data instance and...
2620 (make_pass_sink_code): ...new function.
2621 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2622 subclass of gimple_opt_pass along with...
2623 (pass_data_strlen): ...new pass_data instance and...
2624 (make_pass_strlen): ...new function.
2625 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2626 struct to a subclass of gimple_opt_pass along with...
2627 (pass_data_build_alias): ...new pass_data instance and...
2628 (make_pass_build_alias): ...new function.
2629 (pass_build_ealias): Convert from a global struct to a subclass of
2630 gimple_opt_pass along with...
2631 (pass_data_build_ealias): ...new pass_data instance and...
2632 (make_pass_build_ealias): ...new function.
2633 (pass_ipa_pta): Convert from a global struct to a subclass of
2634 simple_ipa_opt_pass along with...
2635 (pass_data_ipa_pta): ...new pass_data instance and...
2636 (make_pass_ipa_pta): ...new function.
2637 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2638 subclass of gimple_opt_pass along with...
2639 (pass_data_uncprop): ...new pass_data instance and...
2640 (make_pass_uncprop): ...new function.
2641 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2642 global struct to a subclass of gimple_opt_pass along with...
2643 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2644 (make_pass_late_warn_uninitialized): ...new function.
2645 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2646 to a subclass of gimple_opt_pass along with...
2647 (pass_data_init_datastructures): ...new pass_data instance and...
2648 (make_pass_init_datastructures): ...new function.
2649 (pass_early_warn_uninitialized): Convert from a global struct to a
2650 subclass of gimple_opt_pass along with...
2651 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2652 (make_pass_early_warn_uninitialized): ...new function.
2653 (pass_update_address_taken): Convert from a global struct to a
2654 subclass of gimple_opt_pass along with...
2655 (pass_data_update_address_taken): ...new pass_data instance and...
2656 (make_pass_update_address_taken): ...new function.
2657 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2658 struct to a subclass of gimple_opt_pass along with...
2659 (pass_data_release_ssa_names): ...new pass_data instance and...
2660 (make_pass_release_ssa_names): ...new function.
2661 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2662 subclass of gimple_opt_pass along with...
2663 (pass_data_stdarg): ...new pass_data instance and...
2664 (make_pass_stdarg): ...new function.
2665 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2666 global struct to a subclass of gimple_opt_pass along with...
2667 (pass_data_convert_switch): ...new pass_data instance and...
2668 (make_pass_convert_switch): ...new function.
2669 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2670 to a subclass of gimple_opt_pass along with...
2671 (pass_data_tail_recursion): ...new pass_data instance and...
2672 (make_pass_tail_recursion): ...new function.
2673 (pass_tail_calls): Convert from a global struct to a subclass of
2674 gimple_opt_pass along with...
2675 (pass_data_tail_calls): ...new pass_data instance and...
2676 (make_pass_tail_calls): ...new function.
2677 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2678 struct to a subclass of gimple_opt_pass along with...
2679 (pass_data_lower_vector): ...new pass_data instance and...
2680 (make_pass_lower_vector): ...new function.
2681 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2682 gimple_opt_pass along with...
2683 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2684 (make_pass_lower_vector_ssa): ...new function.
2685 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2686 to a subclass of gimple_opt_pass along with...
2687 (pass_data_slp_vectorize): ...new pass_data instance and...
2688 (make_pass_slp_vectorize): ...new function.
2689 (pass_ipa_increase_alignment): Convert from a global struct to a
2690 subclass of simple_ipa_opt_pass along with...
2691 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2692 (make_pass_ipa_increase_alignment): ...new function.
2693 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2694 gimple_opt_pass along with...
2695 (pass_data_vrp): ...new pass_data instance and...
2696 (make_pass_vrp): ...new function.
2697 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2698 subclass of simple_ipa_opt_pass along with...
2699 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2700 (make_pass_ipa_free_lang_data): ...new function.
2701 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2702 gimple_opt_pass along with...
2703 (pass_data_tsan): ...new pass_data instance and...
2704 (make_pass_tsan): ...new function.
2705 (pass_tsan_O0): Convert from a global struct to a subclass of
2706 gimple_opt_pass along with...
2707 (pass_data_tsan_O0): ...new pass_data instance and...
2708 (make_pass_tsan_O0): ...new function.
2709 * var-tracking.c (pass_variable_tracking): Convert from a global
2710 struct to a subclass of rtl_opt_pass along with...
2711 (pass_data_variable_tracking): ...new pass_data instance and...
2712 (make_pass_variable_tracking): ...new function.
2713 * web.c (pass_web): Convert from a global struct to a subclass of
2714 rtl_opt_pass along with...
2715 (pass_data_web): ...new pass_data instance and...
2716 (make_pass_web): ...new function.
2717 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2718 declaration with that of...
2719 (make_pass_mode_switch_use): ...new function.
2720 (pass_resolve_sw_modes): Replace declaration with that of...
2721 (make_pass_resolve_sw_modes): ...new function.
2722 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2723 from a global struct to a subclass of rtl_opt_pass along with...
2724 (pass_data_mode_switch_use): ...new pass_data instance and...
2725 (make_pass_mode_switch_use): ...new function.
2726 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2727 from a global struct to a subclass of rtl_opt_pass along with...
2728 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2729 (make_pass_resolve_sw_modes): ...new function.
2730 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2731 struct to a subclass of rtl_opt_pass along with...
2732 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2733 (make_pass_insert_vzeroupper): ...new function.
2734 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2735 global struct to a subclass of rtl_opt_pass along with...
2736 (pass_data_work_around_errata): ...new pass_data instance and...
2737 (make_pass_work_around_errata): ...new function.
2738 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2739 struct to a subclass of rtl_opt_pass along with...
2740 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2741 (make_pass_mips_machine_reorg2): ...new function.
2743 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2745 * passes.c (pass_manager::operator new): New.
2747 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2749 Handwritten part of conversion of passes to C++ classes.
2751 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2752 (toplev.o): Add dep on PASS_MANAGER_H.
2753 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2754 of early local pases to reflect this moving from a global to a
2755 member of gcc::pass_manager.
2756 (cgraph_add_new_function): Likewise.
2757 * lto-cgraph.c (lto_output_node): Update for conversion of
2758 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2759 * passes.c (opt_pass::clone): New.
2760 (opt_pass::gate): New.
2761 (opt_pass::execute): New.
2762 (opt_pass::opt_pass): New.
2763 (pass_manager::execute_early_local_passes): New.
2764 (pass_manager::execute_pass_mode_switching): new.
2765 (finish_optimization_passes): Convert to...
2766 (pass_manager::finish_optimization_passes): ...this.
2767 (finish_optimization_passes): Update for conversion of passes to
2769 (register_dump_files_1): Use has_gate since we cannot portably
2770 check a vtable entry against NULL.
2771 (dump_one_pass): Likewise.
2772 (ipa_write_summaries_2): Likewise.
2773 (ipa_write_optimization_summaries_1): Likewise.
2774 (ipa_read_summaries_1): Likewise.
2775 (ipa_read_optimization_summaries_1): Likewise.
2776 (execute_ipa_stmt_fixups): Likewise.
2777 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2778 pass-creation functions rather than wiring up globals, and
2779 storing the results in fields of pass_manager generated using
2781 (pass_manager::dump_profile_report): Update for conversion of
2782 passes to C++ classes.
2783 (pass_manager::execute_ipa_summary_passes): Likewise.
2784 (execute_one_ipa_transform_pass): Likewise.
2785 (execute_one_pass): Use has_gate and has_execute since we cannot
2786 portably check a vtable entry against NULL.
2787 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2788 (pass_manager): Use pass-instances.def to add fields for the
2789 various pass instances.
2790 * toplev.c (finalize): Update for move of
2791 finish_optimization_passes to a method of gcc::pass_manager.
2792 * toplev.h (finish_optimization_passes): Move to method of class
2794 * tree-pass.h (struct pass_data): New.
2795 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2796 (opt_pass::gate): Convert to virtual function.
2797 (opt_pass::~opt_pass): New.
2798 (opt_pass::clone): New.
2799 (opt_pass::execute): Convert to virtual function.
2800 (opt_pass::opt_pass): New.
2801 (opt_pass::ctxt_): new.
2802 (gimple_opt_pass): Convert to subclass of opt_pass.
2803 (gimple_opt_pass::gimple_opt_pass): New.
2804 (rtl_opt_pass): Convert to subclass of opt_pass.
2805 (rtl_opt_pass::rtl_opt_pass): New.
2806 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2807 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2808 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2809 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2810 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2811 invocation of pass_mode_switching to reflect this moving from a
2812 global to a member of gcc::pass_manager.
2813 (ix86_option_override): Rework how pass_insert_vzeroupper is
2814 added to the pass_manager to reflect autogenerated changes.
2815 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2817 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2819 PR rtl-optimization/57708
2820 * recog.c (peep2_find_free_register): Validate all regs in a
2823 2013-08-05 Jan Hubicka <jh@suse.cz>
2826 * cgraph.c (verify_cgraph_node): Accept local flags from other
2828 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2829 (function_and_variable_visibility): Likewise.
2830 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2832 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2834 * graph.c (init_graph_slim_pretty_print): Remove.
2835 (print_graph_cfg): Do not call it. Use local pretty printer.
2836 (start_graph_dump): Likewise.
2838 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2840 * gimple-pretty-print.c (buffer): Remove.
2841 (initialized): Likewise.
2842 (maybe_init_pretty_print): Likewise.
2843 (print_gimple_stmt): Do not call it. Use non-static local
2844 pretty_printer variable.
2845 (print_gimple_expr): Likewise.
2846 (print_gimple_seq): Likewise.
2847 (gimple_dump_bb): Likewise.
2849 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2851 * asan.c (asan_pp): Remove.
2852 (asan_pp_initialized): Likewise.
2853 (asan_pp_initialize): Likewise.
2854 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2855 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2856 (asan_add_global): Likewise.
2858 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2860 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2861 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2862 * pretty-print.c (pp_formatted_text_data): Likewise.
2863 (pp_indent): Rename from pp_base_indent.
2864 (pp_format): Rename from pp_base_format.
2865 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2866 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2867 (pp_flush): Rename from pp_base_flush.
2868 (pp_set_line_maximum_length): Rename from
2869 pp_base_set_line_maximum_length.
2870 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2871 (pp_set_prefix): Rename from pp_base_set_prefix.
2872 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2873 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2874 (pp_append_text): Rename from pp_base_append_text.
2875 (pp_formatted_text): Rename from pp_base_formatted_text.
2876 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2877 (pp_remaining_character_count_for_line): Rename from
2878 pp_base_remaining_character_count_for_line.
2879 (pp_newline): Rename from pp_base_newline.
2880 (pp_character): Rename from pp_base_character.
2881 (pp_string): Rename from pp_base_string.
2882 (pp_maybe_space): Rename from pp_base_maybe_space.
2883 * asan.c (asan_pp_string): Adjust.
2884 (asan_emit_stack_protection): Likewise.
2885 (asan_add_global): Likewise.
2886 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2887 * tree-mudflap.c (mf_varname_tree): Likewise.
2888 * tree-pretty-print.c (pp_tree_identifier): Rename from
2889 pp_base_tree_identifier.
2890 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2891 Declare as function.
2893 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2895 * pretty-print.h (pp_bar_bar): New.
2896 (pp_ampersand_ampersand): Likewise.
2897 (pp_less_equal): Likewise.
2898 (pp_greater_equal): Likewise.
2899 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2900 printer functions instead of pp_string or operators and punctuators.
2901 (dump_gimple_call): Likewise.
2902 (dump_gimple_omp_for): Likewise.
2903 (dump_gimple_transaction): Likewise.
2904 (dump_gimple_phi): Likewise.
2905 (pp_gimple_stmt_1): Likewise.
2906 * sched-vis.c (print_insn): Likewise.
2907 * tree-mudflap.c (mf_varname_tree): Likewise.
2908 * tree-pretty-print.c (dump_block_node): Likewise.
2909 (dump_generic_node): Likewise.
2911 2013-08-02 Jan Hubicka <jh@suse.cz>
2913 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2915 * lto-streamer-out.c (tree_is_indexable): Results decls and
2916 parm decls are not indexable.
2917 (DFS_write_tree_body): Do not follow args and results.
2918 (hash_tree): Likewise.
2919 (output_functions): Rearrange so struct function is needed
2920 only when real body is output; be able to also ouptut abstract
2921 functions; output DECL_ARGUMENTS and DECL_RESULT.
2922 (lto_output): When not in WPA, ale store abstract functions.
2923 (write_symbol): Do not care about RESULT_DECL.
2924 (output_symbol_p): Handle correctly sbtract decls.
2925 * lto-streamer-in.c (input_function): Rearrange so struct
2926 function can be NULL at entry; allow streaming of
2927 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2928 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2929 sanity check during LTO.
2930 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2931 RESULT_DECl and DECL_ARGUMENTS.
2932 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2935 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2937 * pretty-print.h (pp_underscore): New.
2939 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2940 printer functions instead of pp_character.
2941 (dump_binary_rhs): Likewise.
2942 (dump_ternary_rhs): Likewise.
2943 (dump_gimple_call_args): Likewise.
2944 (pp_points_to_solution): Likewise.
2945 (dump_gimple_call): Likewise.
2946 (dump_gimple_switch): Likewise.
2947 (dump_gimple_cond): Likewise.
2948 (dump_gimple_bind): Likewise.
2949 (dump_gimple_try): Likewise.
2950 (dump_gimple_omp_for): Likewise.
2951 (dump_gimple_omp_continue): Likewise.
2952 (dump_gimple_omp_single): Likewise.
2953 (dump_gimple_omp_sections): Likewise.
2954 (dump_gimple_omp_block): Likewise.
2955 (dump_gimple_omp_critical): Likewise.
2956 (dump_gimple_transaction): Likewise.
2957 (dump_gimple_asm): Likewise.
2958 (dump_gimple_phi): Likewise.
2959 (dump_gimple_omp_parallel): Likewise.
2960 (dump_gimple_omp_task): Likewise.
2961 (dump_gimple_omp_atomic_load): Likewise.
2962 (dump_gimple_omp_atomic_store): Likewise.
2963 (dump_gimple_mem_ops): Likewise.
2964 (pp_gimple_stmt_1): Likewise.
2965 (pp_cfg_jump): Likewise.
2966 (dump_implicit_edges): Likewise.
2967 (gimple_dump_bb_for_graph): Likewise.
2968 * graph.c (draw_cfg_node): Likewise.
2969 * langhooks.c (lhd_print_error_function): Likewise.
2970 * sched-vis.c (print_exp): Likewise.
2971 (print_value): Likewise.
2972 (print_pattern): Likewise.
2973 (print_insn): Likewise.
2974 (rtl_dump_bb_for_graph): Likewise.
2975 * tree-pretty-print.c (dump_function_declaration): Likewise.
2976 (dump_array_domain): Likewise.
2977 (dump_omp_clause): Likewise.
2978 (dump_location): Likewise.
2979 (dump_generic_node): Likewise.
2980 (print_struct_decl): Likewise.
2981 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2983 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2985 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2986 candidate table when replacing a candidate statement.
2987 (replace_rhs_if_not_dup): Likewise.
2988 (replace_one_candidate): Likewise.
2990 2013-08-02 Jan Hubicka <jh@suse.cz>
2991 Martin Liska <marxin.liska@gmail.com>
2993 * cgraphunit.c (add_new_function): Fix logic when adding from
2995 (assemble_thunk): Rename to ...
2996 (expand_thunk); .. this one; export; get it working with
2997 general functions; make produced gimple valid.
2998 * cgraph.h (expand_thunk): Declare.
3000 2013-08-02 Jan Hubicka <jh@suse.cz>
3002 * ipa-cp.c (gather_context_independent_values): Use
3003 ipa_get_param_move_cost.
3004 (get_replacement_map): Remove PARAM; move parameter folding
3006 (create_specialized_node): Update.
3007 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3008 assert that we have gimple body; update move_cost.
3009 (count_formal_params): Assert that we have gimple body.
3010 (ipa_dump_param): New function.
3011 (ipa_alloc_node_params): Break out from ...
3012 (ipa_initialize_node_params): ... here.
3013 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3014 (ipa_write_node_info): Stream move costs.
3015 (ipa_read_node_info): Read move costs.
3016 (ipa_update_after_lto_read): Do not recompute node params.
3017 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3018 (ipa_get_param): Check we are not in WPA.
3019 (ipa_get_param_move_cost): New.
3020 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3021 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3022 parm numbers to be present.
3024 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3026 PR rtl-optimization/58048
3027 * lra-constraints.c (process_alt_operands): Don't check asm
3028 operand on register.
3030 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3032 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3033 the implied StoreLoad barrier for atomic operations if before.
3035 2013-08-02 Jan Hubicka <jh@suse.cz>
3036 Martin Liska <marxin.liska@gmail.com>
3038 * cgraph.c (cgraph_function_body_availability): Do not check
3040 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3041 symtab_node_availability): Declare.
3042 * ipa.c (can_replace_by_local_alias): New.
3043 (function_and_variable_visibility): Use it.
3044 * symtab.c (symtab_for_node_and_aliases,
3045 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3047 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3049 PR rtl-optimization/57963
3050 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3051 (lra_constraints): Use them.
3053 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3055 * config/arm/types.md (define_attr "type"): Add "load_acq"
3057 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3059 (cortex_a53_store1): Likewise.
3061 2013-08-01 Jan Hubicka <jh@suse.cz>
3063 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3064 partitions are not needed.
3066 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3068 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3069 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3070 MAYBE_NON_Q_CLASS_P where appropriate.
3072 2013-08-01 Jan Hubicka <jh@suse.cz>
3074 * cgraph.h (release_function_body): Declare.
3075 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3076 of unused delcarations.
3078 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3080 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3081 RTL form when subtracting a constant.
3083 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3085 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3086 Generate canonical plus rtx with negated immediate instead of minus
3088 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3090 2013-08-01 Jan Hubicka <jh@suse.cz>
3092 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3093 (cgraph_release_function_body): Likewise.
3094 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3095 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3096 to used_as_abstract_origin.
3097 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3098 symbols used as abstract origins.
3099 * cgraphunit.c (analyze_functions): Update.
3100 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3101 used_as_abstract_origin.
3102 * tree-inline.c (tree_function_versioning): Update
3103 used_as_abstract_origin; be ready for DECL_RESULT and
3104 DECL_ARGUMENTS to be NULL.
3106 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3107 for abstract functions.
3108 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3111 2013-08-01 Jan Hubicka <jh@suse.cz>
3113 * profile.c (compute_value_histograms): Fix thinko.
3115 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3117 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3118 aarch-common-protos.h to extra_headers.
3119 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3120 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3121 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3123 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3125 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3126 Include "../arm/types.md". Define "type" attribute for all patterns.
3127 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3130 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3132 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3133 to support power8 load fusion.
3134 (fusion_gpr_mem_load): Likewise.
3136 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3138 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3139 declarations for power8 load fusion.
3140 (emit_fusion_gpr_load): Likewise.
3142 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3143 tuning for power8, turn on fusion mode by default. Turn on sign
3144 extending fusion mode if normal fusion mode is on, and we are at
3146 (fusion_gpr_load_p): New function, return true if we can fuse an
3147 addis instruction with a dependent load to a GPR.
3148 (emit_fusion_gpr_load): Emit the instructions for power8 load
3151 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3152 (VSX load fusion peepholes): New peepholes to fuse together an
3153 addi instruction with a VSX load instruction.
3155 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3156 peepholes to fuse an addis instruction with a load to a GPR base
3157 register. If we are supporting sign extending fusions, convert
3158 sign extending loads to zero extending loads and add an explicit
3161 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3163 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3164 aarch-common-protos.h to extra_headers.
3165 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3166 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3167 (arm_early_store_addr_dep): Likewise.
3168 (arm_no_early_alu_shift_dep): Likewise.
3169 (arm_no_early_alu_shift_value_dep): Likewise.
3170 (arm_no_early_mul_dep): Likewise.
3171 (arm_no_early_store_addr_dep): Likewise.
3172 (arm_mac_accumulator_is_mul_result): Likewise.
3173 (arm_mac_accumulator_is_result): Likewise.
3174 * config/arm/aarch-common.c: ... here. New file.
3175 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3177 (arm_early_store_addr_dep): Likewise.
3178 (arm_no_early_alu_shift_dep): Likewise.
3179 (arm_no_early_alu_shift_value_dep): Likewise.
3180 (arm_no_early_mul_dep): Likewise.
3181 (arm_no_early_store_addr_dep): Likewise.
3182 (arm_mac_accumulator_is_mul_result): Likewise.
3183 (arm_mac_accumulator_is_result): Likewise.
3184 * config/arm/aarch-common-protos.h: ... here. New file.
3185 * config/arm/t-arm (aarch-common.o): Define.
3187 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3189 * config/arm/arm.md: Include new file "types.md".
3190 (define_attr "type"): Move from here to ...
3191 (define_attr "mul32"): Likewise.
3192 (define_attr "mul64"): Likewise.
3193 * config/arm/types.md: ... here. New file.
3195 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
3197 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
3198 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
3200 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3202 * gen-pass-instances.awk: Fix offset of substr().
3204 2013-07-31 David Malcolm <dmalcolm@redhat.com>
3206 * Makefile.in (pass-instances.def): New.
3207 (passes.o): Replace dependency on passes.def with one on
3210 * gen-pass-instances.awk: New.
3212 * passes.c (pass_manager::pass_manager): Use pass-instances.def
3213 rather than passes.def, updating local definition of NEXT_PASS
3214 macro to add an extra NUM parameter (currently unused).
3216 2013-07-30 David Malcolm <dmalcolm@redhat.com>
3218 * Makefile.in (PASS_MANAGER_H): New.
3219 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
3220 (passes.o): Likewise.
3221 (statistics.o): Likewise.
3222 (cgraphunit.o): Likewise.
3223 (context.o): Depend on PASS_MANAGER_H.
3225 * pass_manager.h: New.
3227 * cgraphunit.c (cgraph_add_new_function): Update for moves
3228 of globals to fields of pass_manager.
3229 (analyze_function): Likewise.
3230 (expand_function): Likewise.
3231 (ipa_passes): Likewise.
3232 (compile): Likewise.
3234 * context.c (context::context): New.
3235 * context.h (context::context): New.
3236 (context::get_passes): New.
3237 (context::passes_): New.
3239 * lto-cgraph.c (input_node): Update for moves of globals to
3240 fields of pass_manager.
3242 * passes.c (all_passes): Remove, in favor of a field of the
3243 same name within the new class pass_manager.
3244 (all_small_ipa_passes): Likewise.
3245 (all_lowering_passes): Likewise.
3246 (all_regular_ipa_passes): Likewise.
3247 (all_late_ipa_passes): Likewise.
3248 (all_lto_gen_passes): Likewise.
3249 (passes_by_id): Likewise.
3250 (passes_by_id_size): Likewise.
3251 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3252 the new class pass_manager.
3253 (set_pass_for_id): Convert to...
3254 (pass_manager::set_pass_for_id): ...method.
3255 (get_pass_for_id): Convert to...
3256 (pass_manager::get_pass_for_id): ...method.
3257 (register_one_dump_file): Move body of implementation into...
3258 (pass_manager::register_one_dump_file): ...here.
3259 (register_dump_files_1): Convert to...
3260 (pass_manager::register_dump_files_1): ...method.
3261 (register_dump_files): Convert to...
3262 (pass_manager::register_dump_files): ...method.
3263 (create_pass_tab): Update for moves of globals to fields of
3265 (dump_passes): Move body of implementation into...
3266 (pass_manager::dump_passes): ...here.
3267 (register_pass): Move body of implementation into...
3268 (pass_manager::register_pass): ...here.
3269 (init_optimization_passes): Convert into...
3270 (pass_manager::pass_manager): ...constructor for new
3271 pass_manager class, and initialize the pass_lists array.
3272 (check_profile_consistency): Update for moves of globals to
3273 fields of pass_manager.
3274 (dump_profile_report): Move body of implementation into...
3275 (pass_manager::dump_profile_report): ...here.
3276 (ipa_write_summaries_1): Update for moves of pass lists from
3277 being globals to fields of pass_manager.
3278 (ipa_write_optimization_summaries): Likewise.
3279 (ipa_read_summaries): Likewise.
3280 (ipa_read_optimization_summaries): Likewise.
3281 (execute_all_ipa_stmt_fixups): Likewise.
3283 * statistics.c (statistics_fini): Update for moves of globals to
3284 fields of pass_manager.
3286 * toplev.c (general_init): Replace call to
3287 init_optimization_passes with construction of the pass_manager
3290 * tree-pass.h (all_passes): Remove, in favor of a field of the
3291 same name within the new class pass_manager.
3292 (all_small_ipa_passes): Likewise.
3293 (all_lowering_passes): Likewise.
3294 (all_regular_ipa_passes): Likewise.
3295 (all_lto_gen_passes): Likewise.
3296 (all_late_ipa_passes): Likewise.
3297 (passes_by_id): Likewise.
3298 (passes_by_id_size): Likewise.
3299 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3300 the new class pass_manager.
3301 (get_pass_for_id): Remove.
3303 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3305 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
3308 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3310 * arm.md (mulhi3): New expand pattern.
3312 2013-07-30 Jan Hubicka <jh@suse.cz>
3313 Martin Liska <marxin.liska@gmail.com>
3315 * profile.c (compute_value_histograms): Do not ICE when
3316 there is mismatch only on some counters.
3318 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3320 PR rtl-optimization/57637
3321 * function.c (move_insn_for_shrink_wrap): Also check the
3322 GEN set of the LIVE problem for the liveness analysis
3323 if it exists, otherwise give up.
3325 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3327 PR tree-optimization/57993
3328 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
3329 replaced statement in the candidate table.
3330 (phi_add_costs): Return infinite cost when the hidden basis does
3331 not dominate all phis on which the candidate is dependent.
3332 (replace_one_candidate): Record replaced statement in the
3335 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
3337 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
3338 (ashlv2si3): New expander.
3339 (*ashlv2si3_i): New define_insn_and_split.
3340 * predicates.md (float_operation): Allow patterns with three
3343 PR rtl-optimization/58021
3344 * mode-switching.c (create_pre_exit): Always split off preceding
3345 insns if we are not at the basic block head.
3347 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
3349 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
3350 (UCLIBC_DYNAMIC_LINKER): New macro.
3351 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
3353 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
3354 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
3356 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
3357 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
3358 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
3359 for SF and DF modes. Use ieee_quad_format for TF mode.
3360 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
3361 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
3362 (mips_option_override): Handle `-mnan=legacy'.
3363 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
3364 `-mabs=2008' and `-mnan=2008'.
3365 (OPTION_DEFAULT_SPECS): Add "nan" default.
3366 (ASM_SPEC): Handle `-mnan='.
3367 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
3368 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
3369 comment accordingly.
3370 (neg<mode>2): Likewise.
3371 * config/mips/mips.opt (mabs, mnan): New options.
3372 * doc/install.texi (Configuration): Document `--with-nan=' option.
3373 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
3375 (MIPS Options): Document them.
3376 * config.gcc <mips*-*-*>: Handle `--with-nan='.
3377 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
3378 * configure: Regenerate.
3379 * config.in: Regenerate.
3381 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3383 * config/i386/i386.md (float post-reload splitters): Do not check
3384 for subregs of SSE registers.
3386 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3387 H.J. Lu <hongjiu.lu@intel.com>
3391 * config/i386/i386.md (post-reload splitter
3392 to avoid partial SSE reg dependency stalls): New pattern.
3394 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
3396 * config/s390/s390.md ("movcc"): Swap load and store instructions.
3398 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
3400 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
3401 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
3403 2013-07-26 Cary Coutant <ccoutant@google.com>
3405 * dwarf2out.c (die_checksum_ordered): Don't include template
3406 instantiations in signature.
3407 (is_template_parameter): New function.
3408 (is_template_instantiation): New function.
3409 (generate_skeleton_bottom_up): Don't include template instantiations
3411 (generate_skeleton): Likewise.
3412 (break_out_comdat_types): Move recursive call to break out nested
3414 (prune_unused_types_mark_generic_parms_dies): Call
3415 is_template_parameter.
3417 2013-07-26 Ian Bolton <ian.bolton@arm.com>
3419 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
3420 uses vector registers.
3421 * config/aarch64/iterators.md: Add attributes rtn and vas.
3423 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3424 Richard Earnshaw <richard.earnshaw@arm.com>
3426 * combine.c (simplify_comparison): Re-canonicalize operands
3428 * config/arm/arm.md (movcond_addsi): New splitter.
3430 2013-07-25 Sterling Augustine <saugustine@google.com>
3432 * dwarf2out.c (size_of_pubnames): Move code to...
3433 (include_pubname_in_output): ...here. New.
3434 (want_pubnames): Rearrange.
3435 (output_pubnames): Call include_pubname_in_output. Move assertion.
3437 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3439 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
3441 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3444 * doc/extend.texi: Remove obsolete builtins. Fix
3445 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
3447 2013-07-25 Jan Hubicka <jh@suse.cz>
3449 * cgraph.c (release_function_body): Break out from ...
3450 (cgraph_release_function_body): ... this one; also release DECL_RESULT
3452 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
3453 old_tree in the map.
3454 (create_specialized_node): Update.
3455 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
3457 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
3458 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
3459 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
3460 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3463 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3465 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3466 addsi3_carryin_alt2_<optab>): Correct output template.
3468 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3470 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3471 Adjust for arm_restrict_it.
3472 Remove trailing whitespace.
3474 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3476 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3477 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3479 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3481 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3483 PR rtl-optimization/57960
3484 * lra-constraints.c (process_alt_operands): Use the right mode
3485 when checking strict_low.
3487 2013-07-25 Jan Hubicka <jh@suse.cz>
3489 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3490 * cgraph.c (cgraph_remove_node): Do not release function body
3491 when in cgraph streaming.
3492 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3493 in other partitions are not considered reachable; fix handling of
3496 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3498 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3500 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3505 * config/arm/arm.md ("*sibcall_insn): Replace use of
3506 Ss with US. Adjust output for v5 and v4t.
3507 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3509 * config/arm/constraints.md ("Ss"): Rename to US.
3511 2013-07-25 Terry Guo <terry.guo@arm.com>
3513 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3514 shift_add/shift_sub0/shift_sub1 RTXs.
3516 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3517 Anton Blanchard <anton@au1.ibm.com>
3519 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3520 (altivec_vpks<VI_char>ss): Likewise.
3521 (altivec_vpks<VI_char>us): Likewise.
3522 (altivec_vpku<VI_char>us): Likewise.
3523 (altivec_vpku<VI_char>um): Likewise.
3525 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3527 Introduce context class.
3529 * Makefile.in (CONTEXT_H): New.
3530 (OBJS): Add context.o.
3531 (toplev.o): Add CONTEXT_H to dependencies.
3534 * toplev.c (general_init): Create the singleton gcc::context instance.
3540 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3542 PR rtl-optimization/57968
3543 * mode-switching.c (create_pre_exit): Allow instructions that
3544 don't set a return register to need a non-exit mode.
3546 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3547 Anton Blanchard <anton@au1.ibm.com>
3549 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3550 operands to vperm for little endian.
3551 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3552 of lvsl to create the control mask for a vperm for little endian.
3554 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3555 Anton Blanchard <anton@au1.ibm.com>
3557 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3558 two operands for little-endian.
3560 2013-07-23 Steve Ellcey <sellcey@mips.com>
3562 * config/mips/mips.c (mips_case_values_threshold): New.
3563 (TARGET_CASE_VALUES_THRESHOLD): Define.
3565 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3566 Anton Blanchard <anton@au1.ibm.com>
3568 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3569 selection of field for vector splat in little endian mode.
3571 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3573 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3574 expanders to rs6000.md.
3575 (ior<mode>3): Likewise.
3576 (and<mode>3): Likewise.
3577 (one_cmpl<mode>2): Likewise.
3578 (nor<mode>3): Likewise.
3579 (andc<mode>3): Likewise.
3580 (eqv<mode>3): Likewise.
3581 (nand<mode>3): Likewise.
3582 (orc<mode>3): Likewise.
3584 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3587 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3588 to split multi-word logical operations.
3589 (rs6000_split_logical_di): Likewise.
3590 (rs6000_split_logical): Likewise.
3592 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3593 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3594 and allow TImode operations in 32-bit.
3595 (vsx_and<mode>3_64bit): Likewise.
3596 (vsx_ior<mode>3_32bit): Likewise.
3597 (vsx_ior<mode>3_64bit): Likewise.
3598 (vsx_xor<mode>3_32bit): Likewise.
3599 (vsx_xor<mode>3_64bit): Likewise.
3600 (vsx_one_cmpl<mode>2_32bit): Likewise.
3601 (vsx_one_cmpl<mode>2_64bit): Likewise.
3602 (vsx_nor<mode>3_32bit): Likewise.
3603 (vsx_nor<mode>3_64bit): Likewise.
3604 (vsx_andc<mode>3_32bit): Likewise.
3605 (vsx_andc<mode>3_64bit): Likewise.
3606 (vsx_eqv<mode>3_32bit): Likewise.
3607 (vsx_eqv<mode>3_64bit): Likewise.
3608 (vsx_nand<mode>3_32bit): Likewise.
3609 (vsx_nand<mode>3_64bit): Likewise.
3610 (vsx_orc<mode>3_32bit): Likewise.
3611 (vsx_orc<mode>3_64bit): Likewise.
3613 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3614 logical types in GPRs.
3616 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3617 logical insns to rs6000.md, and allow TImode operations in
3619 (altivec_ior<mode>3): Likewise.
3620 (altivec_xor<mode>3): Likewise.
3621 (altivec_one_cmpl<mode>2): Likewise.
3622 (altivec_nor<mode>3): Likewise.
3623 (altivec_andc<mode>3): Likewise.
3625 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3626 attributes for moving the 128-bit logical operations into
3628 (BOOL_REGS_OUTPUT): Likewise.
3629 (BOOL_REGS_OP1): Likewise.
3630 (BOOL_REGS_OP2): Likewise.
3631 (BOOL_REGS_UNARY): Likewise.
3632 (BOOL_REGS_AND_CR0): Likewise.
3633 (one_cmpl<mode>2): Add support for DI logical operations on
3634 32-bit, splitting the operations to 32-bit.
3638 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3639 changes to combine the 32/64-bit code, allow logical operations on
3640 TI mode in 32-bit, and to use similar match_operator patterns like
3641 scalar mode uses. Combine the Altivec and VSX code for logical
3642 operations, and move it here.
3643 (ior<mode>3, 128-bit types): Likewise.
3644 (xor<mode>3, 128-bit types): Likewise.
3645 (one_cmpl<mode>3, 128-bit types): Likewise.
3646 (nor<mode>3, 128-bit types): Likewise.
3647 (andc<mode>3, 128-bit types): Likewise.
3648 (eqv<mode>3, 128-bit types): Likewise.
3649 (nand<mode>3, 128-bit types): Likewise.
3650 (orc<mode>3, 128-bit types): Likewise.
3651 (and<mode>3_internal): Likewise.
3652 (bool<mode>3_internal): Likewise.
3653 (boolc<mode>3_internal1): Likewise.
3654 (boolc<mode>3_internal2): Likewise.
3655 (boolcc<mode>3_internal1): Likewise.
3656 (boolcc<mode>3_internal2): Likewise.
3657 (eqv<mode>3_internal1): Likewise.
3658 (eqv<mode>3_internal2): Likewise.
3659 (one_cmpl1<mode>3_internal): Likewise.
3661 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3663 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3664 Rename flag_stack_usage to flag_stack_usage_info.
3666 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3668 * config/microblaze/sync.md: New file.
3669 * config/microblaze/microblaze.md: Include sync.md
3670 * config/microblaze/microblaze.c: Add print_operand 'y'.
3671 * config/microblaze/constraints.md: Add memory_contraint
3672 'Q' which is a single register.
3674 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3676 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3678 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3680 * reload.c (find_reloads): Exit loop once we find this operand
3681 cannot be reloaded somehow for this alternative.
3683 * reload.c (find_reloads): Exit loop once we find a hard register.
3685 * rtlanal.c (computed_jump_p): Exit loop once we find label
3688 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3690 * cfgloopmanip.c (remove_path): Exit loop after setting
3693 * gensupport.c (subst_dup): Avoid loop if code is not
3694 MATCH_DUP nor MATCH_OP_DUP.
3696 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3698 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3700 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3702 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3703 true for SP_REGNUM if mode == ptr_mode.
3704 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3705 with value R0_REGNUM + 31.
3707 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3709 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3710 pad pointer-typed argument downward.
3712 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3714 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3715 and __ILP32__ when the ILP32 model is in use.
3717 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3719 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3720 (aarch64_load_symref_appropriately): In the case of
3721 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3722 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3723 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3724 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3725 if 'mode' doesn't equal to 'ptr_mode'.
3726 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3727 'vcall_offset'; change to call aarch64_emit_move differently depending
3728 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3729 to calculate the upper bound of 'vcall_offset'.
3730 (aarch64_cannot_force_const_mem): Change to also return true if
3732 (aarch64_legitimize_reload_address): In the case of large
3733 displacements, add new local variable 'xmode' and an assertion
3734 based on it; change to use 'xmode' to generate the new rtx and
3736 (aarch64_asm_trampoline_template): Change to generate the template
3737 differently depending on TARGET_ILP32 or not; change to use
3738 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3739 (aarch64_trampoline_size): Removed.
3740 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3741 and replace immediate literals with it. Change to use 'ptr_mode'
3742 instead of 'DImode' and call convert_memory_address if the mode
3743 of 'fnaddr' doesn't equal to 'ptr_mode'.
3744 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3746 (aarch64_elf_asm_destructor): Likewise.
3747 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3748 on TARGET_ILP32 instead of aarch64_trampoline_size.
3749 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3750 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3751 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3752 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3753 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3754 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3755 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3756 depending on the value of 'mode'.
3757 (add_losym_<mode>): New.
3758 (ldr_got_small_<mode>): New, based on ldr_got_small.
3759 (ldr_got_small): Remove.
3760 (ldr_got_small_sidi): New.
3761 * config/aarch64/iterators.md (P): New.
3762 (PTR): Change to 'ptr_mode' in the condition.
3764 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3766 * config.gcc (aarch64*-*-*): Support --with-abi.
3767 (aarch64*-*-elf): Support --with-multilib-list.
3768 (aarch64*-*-linux*): Likewise.
3769 (supported_defaults): Add abi to aarch64*-*-*.
3770 * configure.ac: Mention AArch64 for --with-multilib-list.
3771 * configure: Re-generated.
3772 * config/aarch64/biarchilp32.h: New file.
3773 * config/aarch64/biarchlp64.h: New file.
3774 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3776 (MULTILIB_DEFAULTS): Ditto.
3777 (DRIVER_SELF_SPECS): Ditto.
3778 (ASM_SPEC): Update to also substitute -mabi.
3779 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3780 file whose name depends on -mabi= and -mbig-endian.
3781 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3783 (POINTER_SIZE): New define.
3784 (POINTERS_EXTEND_UNSIGNED): Ditto.
3785 (enum aarch64_abi_type): New enumeration tag.
3786 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3787 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3788 (TARGET_ILP32): New define.
3789 * config/aarch64/aarch64.opt (mabi): New.
3791 (ilp32, lp64): New values for -mabi.
3792 * config/aarch64/t-aarch64 (comma): New define.
3793 (MULTILIB_OPTIONS): Ditto.
3794 (MULTILIB_DIRNAMES): Ditto.
3795 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3796 * doc/invoke.texi: Document -mabi for AArch64.
3798 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3800 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3802 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3803 Anton Blanchard <anton@au1.ibm.com>
3805 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3806 endianness when selecting field to splat.
3808 2013-07-22 Eric Christopher <echristo@gmail.com>
3810 * dwarf2out.c (die_odr_checksum): New function to use
3811 CHECKSUM_ macros and ULEB128 for DIE tag.
3812 (generate_type_signature): Use.
3814 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3816 * config.gcc (sparc*-*-*): Accept leon3 processor.
3817 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3818 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3819 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3820 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3821 (mfix-ut699): Adjust comment.
3822 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3823 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3824 (CPP_CPU_SPEC): Likewise.
3825 (ASM_CPU_SPEC): Likewise.
3826 * config/sparc/sparc.c (leon3_cost): New constant.
3827 (sparc_option_override): Add leon3 support.
3828 (mem_ref): New function.
3829 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3830 (sparc_do_work_around_errata): Look into the instruction in the delay
3831 slot and adjust accordingly. Add fix for the data cache nullify issues
3832 of the UT699. Change insertion position for the NOP.
3833 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3834 (leon3_load): New reservation.
3835 (leon_store): Bump latency to 2.
3836 (grfpu): New automaton.
3837 (grfpu_alu): New unit.
3838 (grfpu_ds): Likewise.
3839 (leon_fp_alu): Adjust.
3840 (leon_fp_mult): Delete.
3841 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3842 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3843 * config/sparc/sparc.md (cpu): Add leon3.
3844 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3846 (atomic_test_and_set): Likewise.
3849 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3851 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3852 default for R5900 targets.
3853 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3854 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3855 * config/mips/mips.c (mips_option_override): Report an error for
3856 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3857 for -march=r5900 -mhard-float.
3859 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3861 * df-problems.c (can_move_insns_across): Exit loop once we
3862 find a non-fixed, non-global register.
3864 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3867 * omega.c (omega_eliminate_red): Break after setting red_found.
3868 (omega_problem_has_red_equations): Similarly after setting found.
3869 (omega_query_variable): Similarly after setting coupled.
3871 2013-07-22 Marek Polacek <polacek@redhat.com>
3873 * gimplify.c: Don't include gimple.h twice.
3875 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3877 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3878 instead of TARGET_THUMB1.
3879 (Pz): New constraint.
3880 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3882 (compare_negsi_si): Likewise.
3883 (compare_addsi2_op0): Likewise.
3884 (compare_addsi2_op1): Likewise.
3885 (addsi3_carryin_<optab>): Likewise.
3886 (addsi3_carryin_alt2_<optab>): Likewise.
3887 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3888 for arm_restrict_it.
3889 (subsi3_carryin): Likewise.
3890 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3891 (minmax_arithsi): Disable for arm_restrict_it.
3892 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3893 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3894 (satsi_<SAT:code>_shift): Likewise.
3895 (arm_shiftsi3): Add alternative for 16-bit encoding.
3896 (arm32_movhf): Disable for arm_restrict_it.
3897 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3898 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3900 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3902 * config/arm/arm.md (attribute "insn"): Delete.
3903 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3904 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3905 (not_shiftsi): Update for attribute change.
3906 (not_shiftsi_compare0): Likewise.
3907 (not_shiftsi_compare0_scratch): Likewise.
3908 (arm_one_cmplsi2): Likewise.
3909 (thumb1_one_cmplsi2): Likewise.
3910 (notsi_compare0): Likewise.
3911 (notsi_compare0_scratch): Likewise.
3912 (thumb1_movdi_insn): Likewise.
3913 (arm_movsi_insn): Likewise.
3914 (movhi_insn_arch4): Likewise.
3915 (movhi_bytes): Likewise.
3916 (arm_movqi_insn): Likewise.
3917 (thumb1_movqi_insn): Likewise.
3918 (arm32_movhf): Likewise.
3919 (thumb1_movhf): Likewise.
3920 (arm_movsf_soft_insn): Likewise.
3921 (thumb1_movsf_insn): Likewise.
3922 (thumb_movdf_insn): Likewise.
3923 (movsicc_insn): Likewise.
3924 (movsfcc_soft_insn): Likewise.
3925 (and_scc): Likewise.
3926 (cond_move): Likewise.
3927 (if_move_not): Likewise.
3928 (if_not_move): Likewise.
3929 (if_shift_move): Likewise.
3930 (if_move_shift): Likewise.
3931 (if_shift_shift): Likewise.
3932 (if_not_arith): Likewise.
3933 (if_arith_not): Likewise.
3934 (cond_move_not): Likewise.
3935 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3936 (neon_mov<mode>): Likewise.
3937 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3938 (thumb2_movsi_vfp): Likewise.
3939 (movsf_vfp): Likewise.
3940 (thumb2_movsf_vfp): Likewise.
3941 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3943 (cortexa7_older_only): Likewise.
3944 (cortexa7_younger): Likewise.
3945 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3946 (1020alu_shift_op): Likewise.
3947 (1020alu_shift_reg_op): Likewise.
3948 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3949 (alu_shift_op): Likewise.
3950 (alu_shift_reg_op): Likewise.
3951 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3952 (11_alu_shift_op): Likewise.
3953 (11_alu_shift_reg_op): Likewise.
3954 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3955 (9_alu_shift_reg_op): Likewise.
3956 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3958 (cortex_a15_alu_shift): Likewise.
3959 (cortex_a15_alu_shift_reg): Likewise.
3960 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3961 (cortex_a5_alu_shift): Likewise.
3962 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3964 (cortex_a53_alu_shift): Likewise.
3965 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3967 (cortex_a7_alu_reg): Likewise.
3968 (cortex_a7_alu_shift): Likewise.
3969 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3970 (cortex_a8_alu_shift): Likewise.
3971 (cortex_a8_alu_shift_reg): Likewise.
3972 (cortex_a8_mov): Likewise.
3973 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3974 (cortex_a9_dp_shift): Likewise.
3975 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3976 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3977 (cortex_r4_mov): Likewise.
3978 (cortex_r4_alu_shift): Likewise.
3979 (cortex_r4_alu_shift_reg): Likewise.
3980 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3981 (526_alu_shift_op): Likewise.
3982 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3983 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3984 (626te_alu_shift_op): Likewise.
3985 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3986 (726te_alu_op): Likewise.
3987 (726te_alu_shift_op): Likewise.
3988 (726te_alu_shift_reg_op): Likewise.
3989 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3990 (mp626_alu_shift_op): Likewise.
3991 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3992 (pj4_alu_e1_conds): Likewise.
3993 (pj4_alu): Likewise.
3994 (pj4_alu_conds): Likewise.
3995 (pj4_shift): Likewise.
3996 (pj4_shift_conds): Likewise.
3997 (pj4_alu_shift): Likewise.
3998 (pj4_alu_shift_conds): Likewise.
4000 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4002 * config/arm/predicates.md (shiftable_operator_strict_it):
4004 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4005 Disable cond_exec version for arm_restrict_it.
4006 (thumb2_smaxsi3): Convert to generate cond_exec.
4007 (thumb2_sminsi3): Likewise.
4008 (thumb32_umaxsi3): Likewise.
4009 (thumb2_uminsi3): Likewise.
4010 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4011 (thumb2_neg_abssi2): Likewise.
4012 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4013 (thumb2_movsicc_insn): Adjust alternatives.
4014 (thumb2_mov_negscc): Disable for arm_restrict_it.
4015 (thumb2_mov_negscc_strict_it): New pattern.
4016 (thumb2_mov_notscc_strict_it): New pattern.
4017 (thumb2_mov_notscc): Disable for arm_restrict_it.
4018 (thumb2_ior_scc): Likewise.
4019 (thumb2_ior_scc_strict_it): New pattern.
4020 (thumb2_cond_move): Adjust for arm_restrict_it.
4021 (thumb2_cond_arith): Disable for arm_restrict_it.
4022 (thumb2_cond_arith_strict_it): New pattern.
4023 (thumb2_cond_sub): Adjust for arm_restrict_it.
4024 (thumb2_movcond): Likewise.
4025 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4026 (thumb2_zero_extendhisi2_v6): Likewise.
4027 (thumb2_zero_extendqisi2_v6): Likewise.
4028 (orsi_notsi_si): Likewise.
4029 (orsi_not_shiftsi_si): Likewise.
4031 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4033 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4034 instruction sequence is 1 byte shorter.
4036 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4038 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4039 it is not needed after split.
4041 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4044 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4045 second label for nonlocal goto receivers. Don't output pic base labels
4046 unless we're producing PIC; mark that action unreachable().
4047 (ix86_save_reg): If the function contains a nonlocal label, save the
4049 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4050 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4051 (update_pic_label_number_if_needed): New.
4052 (machopic_output_function_base_name): Adjust for nonlocal receiver
4054 (machopic_should_output_picbase_label): New.
4055 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4056 (nonlocal_goto_receiver): New insn and split.
4058 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4060 * config/aarch64/aarch64-builtins.c
4061 (aarch64_fold_builtin): Fold abs in all modes.
4062 * config/aarch64/aarch64-simd-builtins.def
4063 (abs): Enable for all modes.
4064 * config/aarch64/arm_neon.h
4065 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4066 (vabs_f64): Add missing intrinsic.
4068 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4070 * config/aarch64/arm_neon.h (vabs_s64): New function
4072 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4075 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4076 * config/avr/avr.md (adjust_len): Add `round'.
4077 * config/avr/avr-protos.h (avr_out_round): New prototype.
4078 (avr_out_plus): Add `out_label' argument.
4079 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4080 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4081 Handle the case where `insn' is just a pattern.
4082 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4083 (avr_out_round): New function.
4084 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4086 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4088 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4089 Add check for flag_stack_usage to handle -fstack-usage support
4091 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4093 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4094 interaction for new Power8 flags and VSX.
4096 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4099 * tree-inline.c (expand_call_inline): Emit errors during
4100 early_inlining only if optimization is not turned on.
4102 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4106 * passes.c (init_optimization_passes): Move the construction of
4107 the pass hierarchy into a new passes.def file.
4109 * Makefile.in (passes.o): Add dependency on passes.def.
4111 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4113 * passes.c (init_optimization_passes): Introduce macros for
4114 constructing the tree of passes (INSERT_PASSES_AFTER,
4115 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4116 TERMINATE_PASS_LIST).
4118 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4119 Wei Mi <wmi@google.com>
4121 PR rtl-optimization/57878
4122 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4124 (reload_pseudo_compare_func): Check nregs first for reload
4127 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4129 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4131 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4133 * read-rtl.c (validate_const_int): Once an invalid character is
4134 seen, quit the loop.
4136 * gengtype.c (write_roots): Similarly once we find the "deletable"
4137 or "if_marked" option.
4139 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4141 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4142 "xtab" and "sat". Move value "clz" from here to ...
4143 (attriubte "type"): ... here.
4144 (satsi_<SAT:code>): Delete "insn" attribute.
4145 (satsi_<SAT:code>_shift): Likewise.
4146 (arm_zero_extendqisi2addsi): Likewise.
4147 (arm_extendqisi2addsi): Likewise.
4148 (clzsi2): Update for attribute changes.
4149 (rbitsi2): Likewise.
4150 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4152 (arm_usatsihi): Likewise.
4153 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4155 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4157 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4158 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4159 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4160 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4161 in alphabetical order.
4162 (attribute "core_cycles"): Update for attribute changes.
4163 (arm_addsi3): Likewise.
4164 (addsi3_compare0): Likewise.
4165 (addsi3_compare0_scratch): Likewise.
4166 (addsi3_compare_op1): Likewise.
4167 (addsi3_compare_op2): Likewise.
4168 (compare_addsi2_op0): Likewise.
4169 (compare_addsi2_op1): Likewise.
4170 (addsi3_carryin_shift_<optab>): Likewise.
4171 (subsi3_carryin_shift): Likewise.
4172 (rsbsi3_carryin_shift): Likewise.
4173 (arm_subsi3_insn): Likewise.
4174 (subsi3_compare0): Likewise.
4175 (subsi3_compare): Likewise.
4176 (arm_andsi3_insn): Likewise.
4177 (thumb1_andsi3_insn): Likewise.
4178 (andsi3_compare0): Likewise.
4179 (andsi3_compare0_scratch): Likewise.
4180 (zeroextractsi_compare0_scratch
4181 (andsi_not_shiftsi_si): Likewise.
4182 (iorsi3_insn): Likewise.
4183 (iorsi3_compare0): Likewise.
4184 (iorsi3_compare0_scratch): Likewise.
4185 (arm_xorsi3): Likewise.
4186 (thumb1_xorsi3_insn): Likewise.
4187 (xorsi3_compare0): Likewise.
4188 (xorsi3_compare0_scratch): Likewise.
4189 (satsi_<SAT:code>_shift): Likewise.
4191 (arm_shiftsi3): Likewise.
4192 (shiftsi3_compare0): Likewise.
4193 (not_shiftsi): Likewise.
4194 (not_shiftsi_compare0): Likewise.
4195 (not_shiftsi_compare0_scratch): Likewise.
4196 (arm_one_cmplsi2): Likewise.
4197 (thumb_one_complsi2): Likewise.
4198 (notsi_compare0): Likewise.
4199 (notsi_compare0_scratch): Likewise.
4200 (thumb1_zero_extendhisi2): Likewise.
4201 (arm_zero_extendhisi2): Likewise.
4202 (arm_zero_extendhisi2_v6): Likewise.
4203 (arm_zero_extendhisi2addsi): Likewise.
4204 (thumb1_zero_extendqisi2): Likewise.
4205 (thumb1_zero_extendqisi2_v6): Likewise.
4206 (arm_zero_extendqisi2): Likewise.
4207 (arm_zero_extendqisi2_v6): Likewise.
4208 (arm_zero_extendqisi2addsi): Likewise.
4209 (thumb1_extendhisi2): Likewise.
4210 (arm_extendhisi2): Likewise.
4211 (arm_extendhisi2_v6): Likewise.
4212 (arm_extendqisi): Likewise.
4213 (arm_extendqisi_v6): Likewise.
4214 (arm_extendqisi2addsi): Likewise.
4215 (thumb1_extendqisi2): Likewise.
4216 (thumb1_movdi_insn): Likewise.
4217 (arm_movsi_insn): Likewise.
4218 (movsi_compare0): Likewise.
4219 (movhi_insn_arch4): Likewise.
4220 (movhi_bytes): Likewise.
4221 (arm_movqi_insn): Likewise.
4222 (thumb1_movqi_insn): Likewise.
4223 (arm32_movhf): Likewise.
4224 (thumb1_movhf): Likewise.
4225 (arm_movsf_soft_insn): Likewise.
4226 (thumb1_movsf_insn): Likewise.
4227 (movdf_soft_insn): Likewise.
4228 (thumb_movdf_insn): Likewise.
4229 (arm_cmpsi_insn): Likewise.
4230 (cmpsi_shiftsi): Likewise.
4231 (cmpsi_shiftsi_swp): Likewise.
4232 (arm_cmpsi_negshiftsi_si): Likewise.
4233 (movsicc_insn): Likewise.
4234 (movsfcc_soft_insn): Likewise.
4235 (arith_shiftsi): Likewise.
4236 (arith_shiftsi_compare0
4237 (arith_shiftsi_compare0_scratch
4238 (sub_shiftsi): Likewise.
4239 (sub_shiftsi_compare0
4240 (sub_shiftsi_compare0_scratch
4241 (and_scc): Likewise.
4242 (cond_move): Likewise.
4243 (if_plus_move): Likewise.
4244 (if_move_plus): Likewise.
4245 (if_move_not): Likewise.
4246 (if_not_move): Likewise.
4247 (if_shift_move): Likewise.
4248 (if_move_shift): Likewise.
4249 (if_shift_shift): Likewise.
4250 (if_not_arith): Likewise.
4251 (if_arith_not): Likewise.
4252 (cond_move_not): Likewise.
4253 (thumb1_ashlsi3): Set type attribute.
4254 (thumb1_ashrsi3): Likewise.
4255 (thumb1_lshrsi3): Likewise.
4256 (thumb1_rotrsi3): Likewise.
4257 (shiftsi3_compare0_scratch): Likewise.
4258 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
4259 (neon_mov<mode>): Likewise.
4260 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
4262 (thumb2_movsi_insn): Likewise.
4263 (thumb2_cmpsi_neg_shiftsi): Likewise.
4264 (thumb2_extendqisi_v6): Likewise.
4265 (thumb2_zero_extendhisi2_v6): Likewise.
4266 (thumb2_zero_extendqisi2_v6): Likewise.
4267 (thumb2_shiftsi3_short): Likewise.
4268 (thumb2_addsi3_compare0_scratch): Likewise.
4269 (orsi_not_shiftsi_si): Likewise.
4270 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4271 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
4273 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4274 (1020alu_shift_op): Likewise.
4275 (1020alu_shift_reg_op): Likewise.
4276 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4277 (alu_shift_op): Likewise.
4278 (alu_shift_reg_op): Likewise.
4279 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4280 (11_alu_shift_op): Likewise.
4281 (11_alu_shift_reg_op): Likewise.
4282 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4283 (9_alu_shift_reg_op): Likewise.
4284 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4286 (cortex_a15_alu_shift): Likewise.
4287 (cortex_a15_alu_shift_reg): Likewise.
4288 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
4290 (cortex_a5_alu_shift): Likewise.
4291 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
4293 (cortex_a53_alu_shift): Likewise.
4294 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4296 (cortex_a7_alu_reg): Likewise.
4297 (cortex_a7_alu_shift): Likewise.
4298 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
4300 (cortex_a8_alu_shift): Likewise.
4301 (cortex_a8_alu_shift_reg): Likewise.
4302 (cortex_a8_mov): Likewise.
4303 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4304 (cortex_a9_dp_shift): Likewise.
4305 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
4307 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
4309 (cortex_r4_mov): Likewise.
4310 (cortex_r4_alu_shift): Likewise.
4311 (cortex_r4_alu_shift_reg): Likewise.
4312 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4313 (526_alu_shift_op): Likewise.
4314 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4315 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4316 (626te_alu_shift_op): Likewise.
4317 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
4318 (726te_alu_op): Likewise.
4319 (726te_alu_shift_op): Likewise.
4320 (726te_alu_shift_reg_op): Likewise.
4321 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4322 (mp626_alu_shift_op): Likewise.
4323 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
4324 (pj4_alu_e1_conds): Likewise.
4325 (pj4_alu): Likewise.
4326 (pj4_alu_conds): Likewise.
4327 (pj4_shift): Likewise.
4328 (pj4_shift_conds): Likewise.
4329 (pj4_alu_shift): Likewise.
4330 (pj4_alu_shift_conds): Likewise.
4331 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4333 (cortexa7_older_only): Likewise.
4334 (cortexa7_younger): Likewise.
4336 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4338 * ipa-pure-const.c (generate_summary): Rename to...
4339 (pure_const_generate_summary): ... this.
4341 2013-07-17 Iain Sandoe <iain@codesourcery.com>
4343 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
4345 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
4348 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
4351 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4353 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
4354 enabled without -march=zEC12.
4355 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
4358 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4360 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
4361 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
4362 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
4363 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
4364 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
4365 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
4366 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
4368 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
4369 and ISA_HAS_NMADD3_NMSUB3.
4370 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
4371 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
4372 (nmsub4<mode>, nmsub3<mode>): Likewise.
4373 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
4375 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4377 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
4378 TARGET_MIPS5400 checking.
4380 2013-07-16 Jakub Jelinek <jakub@redhat.com>
4381 Peter Bergner <bergner@vnet.ibm.com>
4383 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
4384 registers in the comment.
4385 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
4386 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
4387 rather than FIRST_PSEUDO_REGISTERS.
4389 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
4391 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
4392 enable extra ISA flags with TARGET_HTM.
4394 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4396 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
4399 2013-07-15 Cong Hou <congh@google.com>
4401 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
4402 in compare function for sorting.
4404 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
4406 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
4407 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
4408 * config/rs6000/rs6000.opt: Add -mhtm option.
4409 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
4410 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
4411 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
4412 __HTM__ if the HTM instructions are available.
4413 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
4414 htm_spr_reg_operand): New define_predicates.
4415 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
4416 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
4418 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
4419 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
4420 HTM builtin functions.
4421 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
4422 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
4423 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
4424 (rs6000_builtin_mask_calculate): Likewise.
4425 (rs6000_option_override_internal): Likewise.
4426 (bdesc_htm): Add new HTM builtin support.
4427 (htm_spr_num): New function.
4428 (htm_spr_regno): Likewise.
4429 (rs6000_htm_spr_icode): Likewise.
4430 (htm_expand_builtin): Likewise.
4431 (htm_init_builtins): Likewise.
4432 (rs6000_expand_builtin): Add support for HTM builtin functions.
4433 (rs6000_init_builtins): Likewise.
4434 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
4436 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
4437 (TARGET_HTM, MASK_HTM): Define macros.
4438 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
4439 (FIXED_REGISTERS): Likewise.
4440 (CALL_USED_REGISTERS): Likewise.
4441 (CALL_REALLY_USED_REGISTERS): Likewise.
4442 (REG_ALLOC_ORDER): Likewise.
4443 (enum reg_class): Likewise.
4444 (REG_CLASS_NAMES): Likewise.
4445 (REG_CLASS_CONTENTS): Likewise.
4446 (REGISTER_NAMES): Likewise.
4447 (ADDITIONAL_REGISTER_NAMES): Likewise.
4448 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
4449 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
4450 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
4451 * config/rs6000/htm.md: New file.
4452 * config/rs6000/htmintrin.h: New file.
4453 * config/rs6000/htmxlintrin.h: New file.
4455 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
4457 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4458 Define SYMBOL_TINY_GOT, update comment.
4459 * config/aarch64/aarch64.c
4460 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
4461 (aarch64_expand_mov_immediate): Likewise.
4462 (aarch64_print_operand): Likewise.
4463 (aarch64_classify_symbol): Likewise.
4464 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4465 (ldr_got_tiny): Define.
4467 2013-07-13 Tobias Grosser <tobias@grosser.es>
4469 PR tree-optimization/54094
4470 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4471 scheduling dimension for the parallelism check from the polyhedral
4472 information in the AST.
4473 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4476 2013-07-13 Jason Merrill <jason@redhat.com>
4478 * print-tree.c (debug_vec_tree): Use debug_raw.
4479 (debug_raw (vec<tree, va_gc> &)): New.
4480 (debug_raw (vec<tree, va_gc> *)): New.
4481 * tree.h: Declare them.
4483 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4485 * ifcvt.c (ifcvt_after_combine): New static variable.
4486 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4488 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4489 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4490 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4492 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4494 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4496 2013-07-12 Michael Matz <matz@suse.de>
4499 * convert.c (convert_to_real): Reject non-float inner types.
4501 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4503 * config/aarch64/aarch64-protos.h
4504 (aarch64_simd_immediate_valid_for_move): Remove.
4505 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4506 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4508 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4510 2013-07-11 Steve Ellcey <sellcey@mips.com>
4512 * config/mips/mips.c (mips_conditional_register_usage): Do not
4513 use t[0-7] registers in MIPS16 mode when optimizing for size.
4515 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4517 * config/i386/i386.c (dispatch_function_versions): Fix array
4518 indexing of function_version_info to match actual_versions.
4520 2013-07-11 Teresa Johnson <tejohnson@google.com>
4522 * vec.h (struct va_gc): Move release out-of-line.
4523 (va_gc::release): Call ggc_free on released vec.
4525 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4527 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4528 Require GOT register as additional operand in UNSPEC.
4529 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4530 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4531 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4532 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4533 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4534 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4535 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4537 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4540 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4541 name seen by assembler/linker rather if available.
4543 2013-07-11 Andreas Schwab <schwab@suse.de>
4545 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4547 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4549 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4551 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4553 * doc/tm.texi.in: Move hook documentation to ....
4554 * target.def: ... here.
4556 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4558 * doc/tm.texi: Regenerate.
4560 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4563 * doc/invoke.texi: Document -Wconditionally-supported.
4565 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4568 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4571 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4574 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4575 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4576 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4577 Remove duplicate devices.
4578 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4579 * config/avr/t-multilib: Regenerate.
4580 * config/avr/avr-tables.opt: Regenerate.
4581 * doc/avr-mmcu.texi: Regenerate.
4583 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4586 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4588 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4590 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4591 the cost of MULT when optimizing for size.
4593 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4595 * config/cr16/cr16-protos.h: Don't include target.h.
4597 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4599 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4600 adjust register size for TDmode and TFmode for VSX registers.
4602 2013-07-08 Kai Tietz <ktietz@redhat.com>
4605 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4606 hook_bool_const_tree_true.
4608 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4610 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4611 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4612 * config/s390/s390.md: Define FPR*_REGNUM constants.
4613 Fix FPR2_REGNUM constant (18 -> 17).
4614 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4615 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4616 ("trunc<BFP:mode><DFP_ALL:mode>2")
4617 ("trunc<DFP_ALL:mode><BFP:mode>2")
4618 ("*extend<BFP:mode><DFP_ALL:mode>2")
4619 ("*extend<DFP_ALL:mode><BFP:mode>2")
4620 ("extend<BFP:mode><DFP_ALL:mode>2")
4621 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4624 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4626 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4628 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4630 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4631 and cfun_fpr_bit_p to cfun_fpr_save_p.
4632 (s390_frame_area, s390_register_info, s390_frame_info)
4633 (s390_emit_prologue, s390_emit_epilogue)
4634 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4636 * config/s390/s390.h: Define *_REGNUM macros for floating point
4639 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4641 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4643 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4645 PR rtl-optimization/57786
4646 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4647 and break out of the loop when it is set to false.
4649 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4652 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4653 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4654 (const_int 63)) 0)).
4655 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4656 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4657 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4659 PR rtl-optimization/57829
4660 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4661 mask bits outside of mode are just sign-extension from mode to HWI.
4663 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4665 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4666 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4667 adjust_address instead of change_address to keep info about alignment.
4668 (emit_strmov): Remove.
4669 (emit_memmov): New function.
4670 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4671 (expand_movmem_epilogue): Likewise and return updated rtx for
4673 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4674 destination and source.
4675 (decide_alignment): Refactor, handle vector_loop.
4676 (ix86_expand_movmem): Likewise.
4677 (ix86_expand_setmem): Likewise.
4678 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4680 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4682 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4683 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4685 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4687 * config/i386/sse.md (sse_movlhps): Change alternative 3
4688 of operand 2 to "m".
4690 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4693 * config/i386/sse.md (iptr): New mode attribute.
4694 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4695 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4696 (<sse>_vmmul<mode>3): Ditto.
4697 (<sse>_vmdiv<mode>3): Ditto.
4698 (sse_vmrcpv4sf2): Ditto.
4699 (<sse>_vmsqrt<mode>2): Ditto.
4700 (sse_vmrsqrtv4sf2): Ditto.
4701 (<sse>_vm<code><mode>3): Ditto.
4702 (avx_vmcmp<mode>3): Ditto.
4703 (<sse>_vmmaskcmp<mode>3): Ditto.
4704 (<sse>_comi): Ditto.
4705 (<sse>_ucomi): Ditto.
4706 (*xop_vmfrcz_<mode>): Ditto.
4707 (*fmai_fmadd_<mode>): Ditto.
4708 (*fmai_fmsub_<mode>): Ditto.
4709 (*fmai_fnmadd_<mode>): Ditto.
4710 (*fmai_fnmsub_<mode>): Ditto.
4711 (*fma4i_vmfmadd_<mode>): Ditto.
4712 (*fma4i_vmfmsub_<mode>): Ditto.
4713 (*fma4i_vmfnmadd_<mode>): Ditto.
4714 (*fma4i_vmfnmsub_<mode>): Ditto.
4715 (*xop_vmfrcz_<mode>): Ditto.
4716 (sse_cvtps2pi): Ditto.
4717 (sse_cvttps2pi): Ditto.
4718 (sse_cvtss2si): Ditto.
4719 (sse_cvtss2si_2): Ditto.
4720 (sse_cvtss2siq_2): Ditto.
4721 (sse_cvttss2si): Ditto.
4722 (sse_cvttss2siq): Ditto.
4723 (sse_cvtsd2si): Ditto.
4724 (sse_cvtsd2si_2): Ditto.
4725 (sse_cvtsd2siq_2): Ditto.
4726 (sse_cvttsd2si): Ditto.
4727 (sse_cvttsd2siq): Ditto.
4728 (sse_cvtsd2ss): Ditto.
4729 (sse_cvtss2sd): Ditto.
4730 (avx2_pbroadcast<mode>): Ditto.
4731 (avx2_pbroadcast<mode>_1): Ditto.
4732 (*avx_vperm_broadcast_v4sf): Ditto.
4734 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4735 (sse_movlhps): Ditto.
4736 (sse_storehps): Ditto.
4737 (sse_loadhps): Ditto.
4738 (sse_storelps): Ditto.
4739 (sse_loadlps): Ditto.
4740 (*vec_concatv4sf): Ditto.
4741 (*vec_interleave_highv2df): Ditto.
4742 (*vec_interleave_lowv2df): Ditto.
4743 (*vec_extractv2df_1_sse): Ditto.
4744 (*vec_extractv2df_0_sse): Ditto.
4745 (sse2_storelpd): Ditto.
4746 (sse2_loadlpd): Ditto.
4747 (sse2_movsd): Ditto.
4748 (*vec_concatv4si): Ditto.
4749 (vec_concatv2di): Ditto.
4751 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4752 for Intel asm dialect.
4753 (mmx_punpcklwd): Ditto.
4754 (mmx_punpckldq): Ditto.
4756 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4757 for intel assembler dialect.
4759 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4762 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4763 for BUILT_IN_C{LZ,LRSB}*.
4764 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4765 * tree-vrp.c (extract_range_basic): Handle
4766 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4767 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4768 fall thru to code calling set_value*.
4769 * builtins.c (expand_builtin): Remove *IMAX cases.
4770 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4771 if width is bigger than 2*HWI.
4773 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4775 PR rtl-optimization/55342
4776 * lra-int.h (lra_subreg_reload_pseudos): New.
4777 * lra.c: Add undoing optional reloads to the block diagram.
4778 (lra_subreg_reload_pseudos): New.
4779 (lra_optional_reload_pseudos): Change comments.
4780 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4781 lra_optional_reload_pseudos after undo transformations.
4782 * lra-assigns.c (pseudo_prefix_title): New.
4783 (lra_setup_reg_renumber): Use it.
4784 (spill_for): Ditto. Check subreg reload pseudos too.
4785 (assign_by_spills): Consider subreg reload pseudos too.
4786 * lra-constraints.c (simplify_operand_subreg): Use
4787 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4788 (curr_insn_transform): Recognize and do optional reloads.
4789 (undo_optional_reloads): New.
4790 (lra_undo_inheritance): Call undo_optional_reloads.
4792 2013-07-05 Thomas Quinot <quinot@adacore.com>
4794 * tree-complex.c (expand_complex_operations_1): Fix typo.
4796 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4798 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4799 (tune_params): New member 'const vec_costs'.
4800 * config/aarch64/aarch64.c (generic_vector_cost): New.
4801 (generic_tunings): New member 'generic_vector_cost'.
4802 (aarch64_builtin_vectorization_cost): New.
4803 (aarch64_add_stmt_cost): New.
4804 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4805 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4807 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4810 * config/i386/predicates.md (vsib_address_operand): Disallow
4811 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4813 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4816 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4817 expand_builtin_setjmp_receiver.
4818 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4819 with NULL for the label parameter.
4820 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4821 the frame-pointer. Adjust comments.
4822 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4823 only if LABEL is non-NULL.
4825 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4827 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4828 (ARM_ABI_AAPCS64): Ditto.
4830 (ARM_DEFAULT_ABI): Ditto.
4832 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4834 * config/aarch64/aarch64-builtins.c
4835 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4836 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4838 * config/aarch64/aarch64-simd.md
4839 (aarch64_ld1<VALL:mode>): New.
4840 (aarch64_st1<VALL:mode>): Likewise.
4841 * config/aarch64/arm_neon.h
4842 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4844 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4846 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4848 (ix86_option_override_internal):Turn on all -mavx target flags by
4849 default as they are dependent on AVX anyway.
4851 2013-07-02 Cary Coutant <ccoutant@google.com>
4853 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4855 (loc_checksum_ordered): Likewise.
4856 (hash_loc_operands): Remove inline keyword.
4858 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4860 PR tree-optimization/57741
4861 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4862 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4863 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4864 Allow REAL_CST step_exprs if flag_associative_math.
4865 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4867 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4869 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4871 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4873 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4875 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4877 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4879 (iorsi3_insn): Likewise.
4880 (arm_xorsi3): Likewise.
4882 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4884 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4886 (attribute "type"): ... here, and prefix with "wmmx_".
4887 (attribute "core_cycles"): Update for attribute changes.
4888 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4889 (tbcstv4hi): Likewise.
4890 (tbcstv2si): Likewise.
4891 (iwmmxt_iordi3): Likewise.
4892 (iwmmxt_xordi3): Likewise.
4893 (iwmmxt_anddi3): Likewise.
4894 (iwmmxt_nanddi3): Likewise.
4895 (iwmmxt_arm_movdi): Likewise.
4896 (iwmmxt_movsi_insn): Likewise.
4897 (mov<mode>_internal): Likewise.
4898 (and<mode>3_iwmmxt): Likewise.
4899 (ior<mode>3_iwmmxt): Likewise.
4900 (xor<mode>3_iwmmxt): Likewise.
4901 (add<mode>3_iwmmxt): Likewise.
4902 (ssaddv8qi3): Likewise.
4903 (ssaddv4hi3): Likewise.
4904 (ssaddv2si3): Likewise.
4905 (usaddv8qi3): Likewise.
4906 (usaddv4hi3): Likewise.
4907 (usaddv2si3): Likewise.
4908 (sub<mode>3_iwmmxt): Likewise.
4909 (sssubv8qi3): Likewise.
4910 (sssubv4hi3): Likewise.
4911 (sssubv2si3): Likewise.
4912 (ussubv8qi3): Likewise.
4913 (ussubv4hi3): Likewise.
4914 (ussubv2si3): Likewise.
4915 (mulv4hi3_iwmmxt): Likewise.
4916 (smulv4hi3_highpart): Likewise.
4917 (umulv4hi3_highpart): Likewise.
4918 (iwmmxt_wmacs): Likewise.
4919 (iwmmxt_wmacsz): Likewise.
4920 (iwmmxt_wmacu): Likewise.
4921 (iwmmxt_wmacuz): Likewise.
4922 (iwmmxt_clrdi): Likewise.
4923 (iwmmxt_clrv8qi): Likewise.
4924 (iwmmxt_clr4hi): Likewise.
4925 (iwmmxt_clr2si): Likewise.
4926 (iwmmxt_uavgrndv8qi3): Likewise.
4927 (iwmmxt_uavgrndv4hi3): Likewise.
4928 (iwmmxt_uavgv8qi3): Likewise.
4929 (iwmmxt_uavgv4hi3): Likewise.
4930 (iwmmxt_tinsrb): Likewise.
4931 (iwmmxt_tinsrh): Likewise.
4932 (iwmmxt_tinsrw): Likewise.
4933 (iwmmxt_textrmub): Likewise.
4934 (iwmmxt_textrmsb): Likewise.
4935 (iwmmxt_textrmuh): Likewise.
4936 (iwmmxt_textrmsh): Likewise.
4937 (iwmmxt_textrmw): Likewise.
4938 (iwmxxt_wshufh): Likewise.
4939 (eqv8qi3): Likewise.
4940 (eqv4hi3): Likewise.
4941 (eqv2si3): Likewise.
4942 (gtuv8qi3): Likewise.
4943 (gtuv4hi3): Likewise.
4944 (gtuv2si3): Likewise.
4945 (gtv8qi3): Likewise.
4946 (gtv4hi3): Likewise.
4947 (gtv2si3): Likewise.
4948 (smax<mode>3_iwmmxt): Likewise.
4949 (umax<mode>3_iwmmxt): Likewise.
4950 (smin<mode>3_iwmmxt): Likewise.
4951 (umin<mode>3_iwmmxt): Likewise.
4952 (iwmmxt_wpackhss): Likewise.
4953 (iwmmxt_wpackwss): Likewise.
4954 (iwmmxt_wpackdss): Likewise.
4955 (iwmmxt_wpackhus): Likewise.
4956 (iwmmxt_wpackwus): Likewise.
4957 (iwmmxt_wpackdus): Likewise.
4958 (iwmmxt_wunpckihb): Likewise.
4959 (iwmmxt_wunpckihh): Likewise.
4960 (iwmmxt_wunpckihw): Likewise.
4961 (iwmmxt_wunpckilb): Likewise.
4962 (iwmmxt_wunpckilh): Likewise.
4963 (iwmmxt_wunpckilw): Likewise.
4964 (iwmmxt_wunpckehub): Likewise.
4965 (iwmmxt_wunpckehuh): Likewise.
4966 (iwmmxt_wunpckehuw): Likewise.
4967 (iwmmxt_wunpckehsb): Likewise.
4968 (iwmmxt_wunpckehsh): Likewise.
4969 (iwmmxt_wunpckehsw): Likewise.
4970 (iwmmxt_wunpckelub): Likewise.
4971 (iwmmxt_wunpckeluh): Likewise.
4972 (iwmmxt_wunpckeluw): Likewise.
4973 (iwmmxt_wunpckelsb): Likewise.
4974 (iwmmxt_wunpckelsh): Likewise.
4975 (iwmmxt_wunpckelsw): Likewise.
4976 (ror<mode>3): Likewise.
4977 (ashr<mode>3_iwmmxt): Likewise.
4978 (lshr<mode>3_iwmmxt): Likewise.
4979 (ashl<mode>3_iwmmxt): Likewise.
4980 (ror<mode>3_di): Likewise.
4981 (ashr<mode>3_di): Likewise.
4982 (lshr<mode>3_di): Likewise.
4983 (ashl<mode>3_di): Likewise.
4984 (iwmmxt_wmadds): Likewise.
4985 (iwmmxt_wmaddu): Likewise.
4986 (iwmmxt_tmia): Likewise.
4987 (iwmmxt_tmiaph): Likewise.
4988 (iwmmxt_tmiabb): Likewise.
4989 (iwmmxt_tmiatb): Likewise.
4990 (iwmmxt_tmiabt): Likewise.
4991 (iwmmxt_tmiatt): Likewise.
4992 (iwmmxt_tmovmskb): Likewise.
4993 (iwmmxt_tmovmskh): Likewise.
4994 (iwmmxt_tmovmskw): Likewise.
4995 (iwmmxt_waccb): Likewise.
4996 (iwmmxt_wacch): Likewise.
4997 (iwmmxt_waccw): Likewise.
4998 (iwmmxt_waligni): Likewise.
4999 (iwmmxt_walignr): Likewise.
5000 (iwmmxt_walignr0): Likewise.
5001 (iwmmxt_walignr1): Likewise.
5002 (iwmmxt_walignr2): Likewise.
5003 (iwmmxt_walignr3): Likewise.
5004 (iwmmxt_wsadb): Likewise.
5005 (iwmmxt_wsadh): Likewise.
5006 (iwmmxt_wsadbz): Likewise.
5007 (iwmmxt_wsadhz): Likewise.
5008 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5009 (iwmmxt_wabsdiffb): Likewise.
5010 (iwmmxt_wabsdiffh): Likewise.
5011 (iwmmxt_wabsdiffw): Likewise.
5012 (iwmmxt_waddsubhx): Likewise
5013 (iwmmxt_wsubaddhx): Likewise.
5014 (addc<mode>3): Likewise.
5015 (iwmmxt_avg4): Likewise.
5016 (iwmmxt_avg4r): Likewise.
5017 (iwmmxt_wmaddsx): Likewise.
5018 (iwmmxt_wmaddux): Likewise.
5019 (iwmmxt_wmaddsn): Likewise.
5020 (iwmmxt_wmaddun): Likewise.
5021 (iwmmxt_wmulwsm): Likewise.
5022 (iwmmxt_wmulwum): Likewise.
5023 (iwmmxt_wmulsmr): Likewise.
5024 (iwmmxt_wmulumr): Likewise.
5025 (iwmmxt_wmulwsmr): Likewise.
5026 (iwmmxt_wmulwumr): Likewise.
5027 (iwmmxt_wmulwl): Likewise.
5028 (iwmmxt_wqmulm): Likewise.
5029 (iwmmxt_wqmulwm): Likewise.
5030 (iwmmxt_wqmulmr): Likewise.
5031 (iwmmxt_wqmulwmr): Likewise.
5032 (iwmmxt_waddbhusm): Likewise.
5033 (iwmmxt_waddbhusl): Likewise.
5034 (iwmmxt_wqmiabb): Likewise.
5035 (iwmmxt_wqmiabt): Likewise.
5036 (iwmmxt_wqmiatb): Likewise.
5037 (iwmmxt_wqmiatt): Likewise.
5038 (iwmmxt_wqmiabbn): Likewise.
5039 (iwmmxt_wqmiabtn): Likewise.
5040 (iwmmxt_wqmiatbn): Likewise.
5041 (iwmmxt_wqmiattn): Likewise.
5042 (iwmmxt_wmiabb): Likewise.
5043 (iwmmxt_wmiabt): Likewise.
5044 (iwmmxt_wmiatb): Likewise.
5045 (iwmmxt_wmiatt): Likewise.
5046 (iwmmxt_wmiabbn): Likewise.
5047 (iwmmxt_wmiabtn): Likewise.
5048 (iwmmxt_wmiatbn): Likewise.
5049 (iwmmxt_wmiattn): Likewise.
5050 (iwmmxt_wmiawbb): Likewise.
5051 (iwmmxt_wmiawbt): Likewise.
5052 (iwmmxt_wmiawtb): Likewise.
5053 (iwmmxt_wmiawtt): Likewise.
5054 (iwmmxt_wmiawbbn): Likewise.
5055 (iwmmxt_wmiawbtn): Likewise.
5056 (iwmmxt_wmiawtbn): Likewise.
5057 (iwmmxt_wmiawttn): Likewise.
5058 (iwmmxt_wmerge): Likewise.
5059 (iwmmxt_tandc<mode>3): Likewise.
5060 (iwmmxt_torc<mode>3): Likewise.
5061 (iwmmxt_torvsc<mode>3): Likewise.
5062 (iwmmxt_textrc<mode>3): Likewise.
5063 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5064 (wmmxt_pack): Likewise.
5065 (wmmxt_mult_c1): Likewise.
5066 (wmmxt_mult_c2): Likewise.
5067 (wmmxt_alu_c1): Likewise.
5068 (wmmxt_alu_c2): Likewise.
5069 (wmmxt_alu_c3): Likewise.
5070 (wmmxt_transfer_c1): Likewise.
5071 (wmmxt_transfer_c2): Likewise.
5072 (wmmxt_transfer_c3): Likewise.
5073 (marvell_f_iwmmxt_wstr): Likewise.
5074 (marvell_f_iwmmxt_wldr): Likewise.
5076 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5078 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5080 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5083 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5084 * lra-constraints.c (need_for_split_p): Check call used hard regs
5085 living through calls.
5087 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5088 call used regs for call insn.
5090 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5093 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5094 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5096 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5098 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5100 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5102 * lra-constraints.c (need_for_split_p): Check call used hard regs
5103 living through calls.
5105 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5108 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5109 to tie with any other modes. Eliminate Altivec vector mode tests,
5110 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5111 code, to return 0 if testing MODE2 for a condition, if we've
5112 already tested MODE1 for the same condition.
5114 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5116 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5119 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5121 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5122 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5124 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5126 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5128 (aarch64_symbolic_constant_p): Remove.
5129 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5130 static. Fix line length and white space.
5131 (aarch64_symbolic_constant_p): Remove.
5132 * config/aarch64/predicates.md (aarch64_valid_symref):
5133 Use aarch64_classify_symbol_expression.
5135 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5137 * config/arm/constraints.md (Ts): New constraint.
5138 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5140 (compare_scc): Use "Ts" constraint for operand 0.
5141 (ior_scc_scc): Likewise.
5142 (and_scc_scc): Likewise.
5143 (and_scc_scc_nodom): Likewise.
5144 (ior_scc_scc_cmp): Likewise for operand 7.
5145 (and_scc_scc_cmp): Likewise.
5146 * config/arm/thumb2.md (thumb2_movsi_insn):
5147 Add alternatives for 16-bit encodings.
5148 (thumb2_movhi_insn): Likewise.
5149 (thumb2_movsicc_insn): Likewise.
5150 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5151 (thumb2_negscc): Use "Ts" constraint.
5152 Move mvn instruction outside cond_exec block.
5153 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5154 for 16-bit encodings.
5156 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5158 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5160 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5161 (mulsi3subsi): Likewise.
5162 (mulsidi3adddi): Likewise.
5163 (mulsidi3_v6): Likewise.
5164 (umulsidi3_v6): Likewise.
5165 (umulsidi3adddi_v6): Likewise.
5166 (smulsi3_highpart_v6): Likewise.
5167 (umulsi3_highpart_v6): Likewise.
5168 (mulhisi3tb): Likewise.
5169 (mulhisi3bt): Likewise.
5170 (mulhisi3tt): Likewise.
5171 (maddhisi4): Likewise.
5172 (maddhisi4tb): Likewise.
5173 (maddhisi4tt): Likewise.
5174 (maddhidi4): Likewise.
5175 (maddhidi4tb): Likewise.
5176 (maddhidi4tt): Likewise.
5177 (zeroextractsi_compare0_scratch): Likewise.
5178 (insv_zero): Likewise.
5179 (insv_t2): Likewise.
5180 (anddi_notzesidi_di): Likewise.
5181 (anddi_notsesidi_di): Likewise.
5182 (andsi_notsi_si): Likewise.
5183 (iordi_zesidi_di): Likewise.
5184 (xordi_zesidi_di): Likewise.
5185 (andsi_iorsi3_notsi): Likewise.
5187 (smax_m1): Likewise.
5189 (not_shiftsi): Likewise.
5190 (unaligned_loadsi): Likewise.
5191 (unaligned_loadhis): Likewise.
5192 (unaligned_loadhiu): Likewise.
5193 (unaligned_storesi): Likewise.
5194 (unaligned_storehi): Likewise.
5195 (extv_reg): Likewise.
5196 (extzv_t2): Likewise.
5198 (udivsi3): Likewise.
5199 (arm_zero_extendhisi2addsi): Likewise.
5200 (arm_zero_extendqisi2addsi): Likewise.
5201 (compareqi_eq0): Likewise.
5202 (arm_extendhisi2_v6): Likewise.
5203 (arm_extendqisi2addsi): Likewise.
5204 (arm_movt): Likewise.
5205 (thumb2_ldrd): Likewise.
5206 (thumb2_ldrd_base): Likewise.
5207 (thumb2_ldrd_base_neg): Likewise.
5208 (thumb2_strd): Likewise.
5209 (thumb2_strd_base): Likewise.
5210 (thumb2_strd_base_neg): Likewise.
5211 (arm_negsi2): Add alternative for 16-bit encoding.
5212 (arm_one_cmplsi2): Likewise.
5214 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5216 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
5217 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
5218 (movdfcc): Likewise.
5219 * config/arm/vfp.md (*thumb2_movsf_vfp):
5220 Disable predication for arm_restrict_it.
5221 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
5222 (*thumb2_movdfcc_vfp): Likewise.
5223 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
5224 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
5225 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
5226 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
5227 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
5228 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
5229 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
5230 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
5231 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
5232 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
5233 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
5234 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
5235 Disable predication for arm_restrict_it.
5237 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
5239 * config/i386/bmiintrin.h (_bextr_u32): New.
5240 (_bextr_u64): Ditto.
5242 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
5244 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
5245 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
5246 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
5247 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
5248 * config/mips/n32-elf.h: ...this new file.
5250 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
5253 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
5254 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
5256 2013-06-27 Catherine Moore <clm@codesourcery.com>
5258 * config/mips/mips-tables.opt: Regenerate.
5259 * config/mips/mips-cpus.def: Add m14ke and m14kec.
5260 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
5261 * doc/invoke.texi: Add -m14kc.
5263 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5266 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
5267 constraints of operand 1 and 2.
5270 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
5271 to match RTL canonicalization. Swap predicates and
5272 constraints of operand 1 and 2.
5274 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
5276 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
5277 Process OP_INOUT regs for splitting too.
5279 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5281 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
5282 decl before the loop, initialize to NULL.
5283 (vectorizable_load): Initialize ptr_incr to NULL.
5285 2013-06-27 Martin Jambor <mjambor@suse.cz>
5288 * ipa-ref.h (ipa_maybe_record_reference): Declare.
5289 * ipa-ref.c (ipa_maybe_record_reference): New function.
5290 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
5291 * ipa-cp.c (create_specialized_node): Record potential references from
5293 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
5295 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5297 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
5298 parameter 'mode' of type 'enum machine_mode mode'; change to pass
5299 'mode' to force_reg.
5300 (aarch64_add_offset): Update calls to aarch64_force_temporary.
5301 (aarch64_expand_mov_immediate): Likewise.
5303 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5305 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
5306 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
5308 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5310 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
5311 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5312 (struct machine_function): Add tbegin_p.
5313 (s390_canonicalize_comparison): Fold CC mode compares to
5314 conditional jump if possible.
5315 (s390_emit_jump): Return the emitted jump.
5316 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
5317 Handle CCRAWmode compares.
5318 (s390_option_override): Default to -mhtm if available.
5319 (s390_reg_clobbered_rtx): Handle floating point regs as well.
5320 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
5321 FPRs instead of df_regs_ever_live_p.
5322 (s390_optimize_nonescaping_tx): New function.
5323 (s390_init_frame_layout): Extend clobbered_regs array to cover
5325 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
5326 (s390_expand_tbegin): New function.
5327 (enum s390_builtin): New enum definition.
5328 (code_for_builtin): New array definition.
5329 (s390_init_builtins): New function.
5330 (s390_expand_builtin): New function.
5331 (TARGET_INIT_BUILTINS): Define.
5332 (TARGET_EXPAND_BUILTIN): Define.
5333 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5334 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
5335 (s390_alc_comparison): Likewise.
5336 * config/s390/s390-modes.def: Add CCRAWmode.
5337 * config/s390/s390.h (processor_flags): Add PF_TX.
5338 (TARGET_CPU_HTM): Define macro.
5339 (TARGET_HTM): Define macro.
5340 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
5341 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
5342 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5343 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
5344 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5345 (TBEGIN_MASK, TBEGINC_MASK): New constants.
5346 ("*cc_to_int"): Move up.
5347 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
5348 constants other than 0.
5349 ("*ccraw_to_int"): New insn and splitter definition.
5350 ("tbegin", "tbegin_nofloat", "tbegin_retry")
5351 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
5352 ("tx_assist"): New expander.
5353 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
5354 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
5355 * config/s390/s390.opt: Add -mhtm option.
5356 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
5357 * config/s390/htmxlintrin.h: New file.
5358 * config/s390/htmintrin.h: New file.
5359 * config/s390/s390intrin.h: New file.
5360 * doc/extend.texi: Document htm builtins.
5361 * config.gcc: Add the new header files to extra_headers.
5363 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
5365 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
5366 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
5368 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
5369 Pat Haugen <pthaugen@us.ibm.com>
5370 Peter Bergner <bergner@vnet.ibm.com>
5372 * config/rs6000/power8.md: New.
5373 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
5374 setting for power8 entry.
5375 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
5376 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
5377 test for Power4/Power5 only.
5378 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
5380 (force_new_group): Adjust comment.
5381 * config/rs6000/rs6000.md: Include power8.md.
5383 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
5385 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
5386 * config/arm/arm-protos.h (arm_max_conditional_execute): New
5388 (tune_params): Update comment.
5389 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
5390 (arm_max_conditional_execute): New function.
5391 (thumb2_final_prescan_insn): Use max_insn_skipped and
5392 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
5394 2013-06-25 Jakub Jelinek <jakub@redhat.com>
5396 PR tree-optimization/57705
5397 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
5398 SSA_NAME step, provided that it is not defined inside the loop.
5399 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
5400 (get_initial_def_for_induction): Handle SSA_NAME IV step.
5402 2013-06-25 Martin Jambor <mjambor@suse.cz>
5405 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
5406 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
5408 (ipa_note_param_call): Initialize member_ptr flag.
5409 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
5410 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
5411 (ipa_write_indirect_edge_info): Stream member_ptr flag.
5412 (ipa_read_indirect_edge_info): Likewise.
5414 2013-06-25 Richard Biener <rguenther@suse.de>
5417 * passes.c (init_optimization_passes): Move pass_fold_builtins
5418 and pass_dce earlier with -Og.
5420 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
5422 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
5423 <BIT_FIELD_REF>: Remove trailing TAB.
5424 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
5427 2013-06-24 Martin Jambor <mjambor@suse.cz>
5429 PR tree-optimization/57358
5430 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
5431 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
5432 (ipa_analyze_params_uses): Generate pessimistic info when true.
5434 2013-06-24 Martin Jambor <mjambor@suse.cz>
5436 PR tree-optimization/57539
5437 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
5438 global.inlined_to of the new node to it. All callers changed.
5439 * ipa-inline-transform.c (clone_inlined_nodes): New variable
5440 inlining_into, pass it to cgraph_clone_node.
5441 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
5442 ipa_free_edge_args_substructures.
5443 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
5444 rdesc linked list. Do not assert rdesc edges have inlined caller.
5445 Assert we have found an rdesc in the rdesc list.
5447 2013-06-24 Richard Biener <rguenther@suse.de>
5449 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
5450 (pointer_set_lookup): Declare.
5451 (class pointer_map): New template class implementing a
5452 generic pointer to T map.
5453 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
5454 pointer_map<T>::contains, pointer_map<T>::insert,
5455 pointer_map<T>::traverse): New functions.
5456 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
5457 (pointer_set_lookup): New function.
5458 (pointer_set_contains): Use pointer_set_lookup.
5459 (pointer_set_insert): Likewise.
5460 (insert_aux): Remove.
5461 (struct pointer_map_t): Embed a pointer_set_t.
5462 (pointer_map_create): Adjust.
5463 (pointer_map_destroy): Likewise.
5464 (pointer_map_contains): Likewise.
5465 (pointer_map_insert): Likewise.
5466 (pointer_map_traverse): Likewise.
5467 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5468 pointer_map<unsigned> instead of a pointer_map_t.
5469 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5470 (streamer_tree_cache_lookup): Likewise.
5471 (streamer_tree_cache_create): Likewise.
5472 (streamer_tree_cache_delete): Likewise.
5473 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5474 pointer_map<unsigned> instead of a pointer_map_t.
5475 (lto_init_tree_ref_encoder): Adjust.
5476 (lto_destroy_tree_ref_encoder): Likewise.
5477 * lto-section-out.c (lto_output_decl_index): Likewise.
5478 (lto_record_function_out_decl_state): Likewise.
5479 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5481 2013-06-24 Richard Biener <rguenther@suse.de>
5483 PR tree-optimization/57488
5484 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5486 2013-06-24 Alan Modra <amodra@gmail.com>
5488 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5489 (gen_easy_altivec_constant): Likewise.
5490 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5491 easy_vector_constant_msb): Likewise.
5493 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5496 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5497 add missing return true.
5499 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5502 * config/sh/predicates.md (general_extend_operand): Invoke
5503 general_movsrc_operand for memory operands.
5504 (general_movsrc_operand): Allow reg+reg addressing, do not use
5505 general_operand for memory operands.
5507 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5509 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5510 when current target options does not apply.
5511 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5512 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5513 * config/i386/bmiintrin.h: Pass appropriate target
5514 attributes to header.
5515 * config/i386/mmintrin.h: Ditto.
5516 * config/i386/nmmintrin.h: Ditto.
5517 * config/i386/avx2intrin.h: Ditto.
5518 * config/i386/fxsrintrin.h: Ditto.
5519 * config/i386/tbmintrin.h: Ditto.
5520 * config/i386/xsaveintrin.h: Ditto.
5521 * config/i386/f16cintrin.h: Ditto.
5522 * config/i386/xtestintrin.h: Ditto.
5523 * config/i386/xsaveoptintrin.h: Ditto.
5524 * config/i386/bmi2intrin.h: Ditto.
5525 * config/i386/lzcntintrin.h: Ditto.
5526 * config/i386/smmintrin.h: Ditto.
5527 * config/i386/wmmintrin.h: Ditto.
5528 * config/i386/x86intrin.h: Remove all header include guards.
5529 * config/i386/prfchwintrin.h: Ditto.
5530 * config/i386/pmmintrin.h: Ditto.
5531 * config/i386/tmmintrin.h: Ditto.
5532 * config/i386/xmmintrin.h: Ditto.
5533 * config/i386/popcntintrin.h: Ditto.
5534 * config/i386/rdseedintrin.h: Ditto.
5535 * config/i386/ammintrin.h: Ditto.
5536 * config/i386/emmintrin.h: Ditto.
5537 * config/i386/immintrin.h: Remove all header include guards.
5538 * config/i386/fma4intrin.h: Ditto.
5539 * config/i386/lwpintrin.h: Ditto.
5540 * config/i386/xopintrin.h: Ditto.
5541 * config/i386/ia32intrin.h: Ditto.
5542 * config/i386/avxintrin.h: Ditto.
5543 * config/i386/rtmintrin.h: Ditto.
5544 * config/i386/fmaintrin.h: Ditto.
5545 * config/i386/mm3dnow.h: Ditto.
5547 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5549 * common/config/i386/i386-common.c: Handle LZCNT.
5551 2013-06-22 Andi Kleen <ak@linux.intel.com>
5553 * doc/extend.texi: Use __atomic_store_n instead of
5554 __atomic_store in HLE example.
5556 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5558 * config/sh/sh.c: Remove <cstdlib> workaround.
5560 2013-06-21 Andi Kleen <ak@linux.intel.com>
5562 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5564 2013-06-21 Andi Kleen <ak@linux.intel.com>
5566 * doc/extend.texi: Document that __atomic_clear and
5567 __atomic_test_and_set should only be used with bool.
5569 2013-06-20 Jan Hubicka <jh@suse.cz>
5571 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5573 * tree.c (decls_same_for_odr): New function.
5574 (same_for_edr): New function.
5575 (types_same_for_odr): New function.
5576 (get_binfo_at_offset): Use it.
5577 * tree.h (types_same_for_odr): Declare.
5579 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5580 Jason Merrill <jason@redhat.com>
5582 * system.h: Include <cstdlib> as well as <stdlib.h>.
5584 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5587 * config/i386/i386.c (construct_container): Report error if
5588 long double is used with disabled x87 float returns.
5590 2013-06-20 Jan Hubicka <jh@suse.cz>
5592 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5594 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5596 PR rtl-optimization/57425
5597 PR rtl-optimization/57569
5598 * alias.c (write_dependence_p): Remove parameters mem_mode and
5599 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5600 Changed all callers.
5601 (canon_anti_dependence): Get comments and semantics in sync.
5602 Add parameter mem_canonicalized. Changed all callers.
5603 * rtl.h (canon_anti_dependence): Update prototype.
5605 2013-06-20 Richard Biener <rguenther@suse.de>
5607 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5608 byte case, inline streamer_read_uchar and defer section
5611 2013-06-20 Richard Biener <rguenther@suse.de>
5613 PR tree-optimization/57584
5614 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5615 SSA names into the expanded expression that take part in
5616 abnormal coalescing.
5618 2013-06-19 Sharad Singhai <singhai@google.com>
5620 * gcov.c (print_usage): Handle new option.
5621 (process_args): Ditto.
5622 (get_gcov_intermediate_filename): New function.
5623 (output_intermediate_file): New function.
5624 (output_gcov_file): New function
5625 (generate_results): Handle new option.
5626 (release_function): Relase demangled name.
5627 (read_graph_file): Handle demangled name.
5628 (output_lines): Ditto.
5629 * doc/gcov.texi: Document gcov intermediate format.
5631 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5634 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5635 (lra_emit_add): Use the functions. Add comment about Y as an
5638 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5641 * collect2.c (collect_atexit): New.
5642 (collect_exit): Delete.
5643 (main): Register collect_atexit with atexit.
5644 (collect_wait): Change collect_exit to exit.
5646 * collect2.h (collect_exit): Delete.
5647 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5649 2013-06-19 Wei Mi <wmi@google.com>
5651 PR rtl-optimization/57518
5652 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5653 if regno is used in paradoxical subreg.
5654 (update_equiv_regs): Check pdx_subregs[regno] before
5655 set a reg to be equivalent with a mem.
5657 2013-06-19 Matthias Klose <doko@ubuntu.com>
5660 * file-find.h (find_a_file): Add a mode parameter.
5661 * file-find.c (find_a_file): Likewise.
5662 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5663 with X_OK for the executables.
5664 * collect2.c (main): Call find_a_file with X_OK.
5666 2013-06-19 Steve Ellcey <sellcey@mips.com>
5669 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5670 Use NEXT_INSN instead of next_real_insn.
5672 2013-06-19 Jan Hubicka <jh@suse.cz>
5674 * cgraph.h (const_value_known_p): Replace by ...
5675 (ctor_for_folding): .. this one.
5676 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5677 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5678 * expr.c (expand_expr_real_1): Likewise.
5679 (string_constant): Likewise.
5680 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5681 * ipa.c (process_references): Likewise.
5682 (symtab_remove_unreachable_nodes): Likewise.
5683 * ipa-inline-analysis.c (param_change_prob): Likewise.
5684 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5685 (get_base_constructor): Likwise.
5686 * varpool.c (varpool_remove_node): Likewise.
5687 (varpool_remove_initializer): LIkewise.
5688 (dump_varpool_node): LIkwise.
5689 (const_value_known_p): Rewrite to ...
5690 (ctor_for_folding): ... this one.
5692 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5695 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5696 PERSONALITY in $PATH derived prefixes.
5698 2013-06-19 Jeff Law <law@redhat.com>
5700 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5703 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5704 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5707 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5709 * config/arm/vfp.md: Move VFP instruction classification documentation
5711 * config/arm/arm.md: ... here. Update instruction classification
5714 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5716 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5717 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5718 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5720 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5723 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5724 with NEXT_INSN. Use tablejump_p to check for jump table data
5727 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5730 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5731 that now in C++ the value is correct per the C++ standards.
5733 2013-06-19 Richard Biener <rguenther@suse.de>
5735 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5738 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5741 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5744 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5745 with next_active_insn.
5747 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5749 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5750 functions are inlined during failures to flag an error.
5751 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5752 in early inline pass.
5754 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5756 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5759 2013-06-18 Julian Brown <julian@codesourcery.com>
5761 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5762 Permit virtual register pre-reload if !strict.
5763 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5765 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5767 * config/arm/neon.md (movmisalign<mode>): Use
5768 neon_perm_struct_or_reg_operand instead of
5769 neon_struct_or_register_operand.
5770 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5771 neon_permissive_struct_operand instead of neon_struct_operand.
5772 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5773 neon_vector_mem_operand.
5774 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5775 neon_vector_mem_operand.
5776 (neon_permissive_struct_operand): New.
5777 (neon_struct_or_register_operand): Rename to...
5778 (neon_perm_struct_or_reg_operand): This. Adjust call to
5779 neon_vector_mem_operand.
5781 2013-06-18 Richard Biener <rguenther@suse.de>
5783 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5784 * lto-streamer.h: Include pointer-set.h.
5785 (struct lto_decl_slot): Remove.
5786 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5787 Remove next_index entry.
5788 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5789 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5790 (lto_init_tree_ref_encoder): Adjust.
5791 (lto_destroy_tree_ref_encoder): Likewise.
5792 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5793 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5794 (lto_output_decl_index): Adjust.
5795 (lto_new_out_decl_state): Likewise.
5796 (lto_record_function_out_decl_state): Likewise.
5797 * lto-streamer-out.c (copy_function): Likewise.
5799 2013-06-18 Richard Biener <rguenther@suse.de>
5801 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5802 * cgraphunit.c: Include cfgloop.h.
5803 (init_lowered_empty_function): Initialize the loop tree.
5804 (assemble_thunk): Insert new BBs into loops.
5806 2013-06-18 Richard Biener <rguenther@suse.de>
5808 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5809 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5810 the map from cache entry to cache index optional.
5811 (streamer_tree_cache_replace_tree): Adjust accordingly.
5812 (streamer_tree_cache_append): Likewise.
5813 (streamer_tree_cache_delete): Likewise.
5814 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5815 streamer cache map from cache entry to cache index.
5816 * lto-streamer-out.c (create_output_block): Adjust.
5818 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5820 * config/arm/arm.md (attribute "insn"): Move multiplication and
5821 division attributes to...
5822 (attribute "type"): ... here. Remove mult.
5823 (attribute "mul32"): New attribute.
5824 (attribute "mul64"): Add umaal.
5825 (*arm_mulsi3): Update attributes.
5826 (*arm_mulsi3_v6): Likewise.
5827 (*thumb_mulsi3): Likewise.
5828 (*thumb_mulsi3_v6): Likewise.
5829 (*mulsi3_compare0): Likewise.
5830 (*mulsi3_compare0_v6): Likewise.
5831 (*mulsi_compare0_scratch): Likewise.
5832 (*mulsi_compare0_scratch_v6): Likewise.
5833 (*mulsi3addsi): Likewise.
5834 (*mulsi3addsi_v6): Likewise.
5835 (*mulsi3addsi_compare0): Likewise.
5836 (*mulsi3addsi_compare0_v6): Likewise.
5837 (*mulsi3addsi_compare0_scratch): Likewise.
5838 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5839 (*mulsi3subsi): Likewise.
5840 (*mulsidi3adddi): Likewise.
5841 (*mulsi3addsi_v6): Likewise.
5842 (*mulsidi3adddi_v6): Likewise.
5843 (*mulsidi3_nov6): Likewise.
5844 (*mulsidi3_v6): Likewise.
5845 (*umulsidi3_nov6): Likewise.
5846 (*umulsidi3_v6): Likewise.
5847 (*umulsidi3adddi): Likewise.
5848 (*umulsidi3adddi_v6): Likewise.
5849 (*smulsi3_highpart_nov6): Likewise.
5850 (*smulsi3_highpart_v6): Likewise.
5851 (*umulsi3_highpart_nov6): Likewise.
5852 (*umulsi3_highpart_v6): Likewise.
5853 (mulhisi3): Likewise.
5854 (*mulhisi3tb): Likewise.
5855 (*mulhisi3bt): Likewise.
5856 (*mulhisi3tt): Likewise.
5857 (maddhisi4): Likewise.
5858 (*maddhisi4tb): Likewise.
5859 (*maddhisi4tt): Likewise.
5860 (maddhidi4): Likewise.
5861 (*maddhidi4tb): Likewise.
5862 (*maddhidi4tt): Likewise.
5864 (udivsi3): Likewise.
5865 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5866 (thumb2_mulsi_short_compare0): Likewise.
5867 (thumb2_mulsi_short_compare0_scratch): Likewise.
5868 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5869 (1020mult2): Likewise.
5870 (1020mult3): Likewise.
5871 (1020mult4): Likewise.
5872 (1020mult5): Likewise.
5873 (1020mult6): Likewise.
5874 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5876 (cortex_a15_mult64): Likewise.
5877 (cortex_a15_sdiv): Likewise.
5878 (cortex_a15_udiv): Likewise.
5879 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5885 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5886 (pj4_ir_div): Likewise.
5887 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5888 (11_mult2): Likewise.
5889 (11_mult3): Likewise.
5890 (11_mult4): Likewise.
5891 (11_mult5): Likewise.
5892 (11_mult6): Likewise.
5893 (11_mult7): Likewise.
5894 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5895 (cortex_a8_mla): Likewise.
5896 (cortex_a8_mull): Likewise.
5897 (cortex_a8_smulwy): Likewise.
5898 (cortex_a8_smlald): Likewise.
5899 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5900 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5901 (cortex_r4_mul_3): Likewise.
5902 (cortex_r4_mla_4): Likewise.
5903 (cortex_r4_mla_3): Likewise.
5904 (cortex_r4_smlald): Likewise.
5905 (cortex_r4_mull): Likewise.
5906 (cortex_r4_sdiv): Likewise.
5907 (cortex_r4_udiv): Likewise.
5908 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5909 (cortex_a7_idiv): Likewise.
5910 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5911 (9_mult2): Likewise.
5912 (9_mult3): Likewise.
5913 (9_mult4): Likewise.
5914 (9_mult5): Likewise.
5915 (9_mult6): Likewise.
5916 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5917 (cortex_a53_sdiv): Likewise.
5918 (cortex_a53_udiv): Likewise.
5919 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5920 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5921 (mp626_mult2): Likewise.
5922 (mp626_mult3): Likewise.
5923 (mp626_mult4): Likewise.
5924 * config/arm/fa526.md (526_mult1): Update attribute change.
5925 (526_mult2): Likewise.
5926 * config/arm/arm-generic.md (mult): Update attribute change.
5927 (mult_ldsched_strongarm): Likewise.
5928 (mult_ldsched): Likewise.
5929 (multi_cycle): Likewise.
5930 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5931 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5932 (606te_mult2): Likewise.
5933 (606te_mult3): Likewise.
5934 (606te_mult4): Likewise.
5935 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5936 (cortex_a9_mac16): Likewise.
5937 (cortex_a9_multiply): Likewise.
5938 (cortex_a9_mac): Likewise.
5939 (cortex_a9_multiply_long): Likewise.
5940 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5941 (626te_mult2): Likewise.
5942 (626te_mult3): Likewise.
5943 (626te_mult4): Likewise.
5945 2013-06-18 Richard Biener <rguenther@suse.de>
5948 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5950 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5953 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5954 with next_active_insn.
5956 2013-06-18 Alan Modra <amodra@gmail.com>
5958 * config/rs6000/rs6000.h (enum data_align): New.
5959 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5960 (DATA_ABI_ALIGNMENT): Define.
5961 (CONSTANT_ALIGNMENT): Correct comment.
5962 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5963 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5965 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5967 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5968 ATTRIBUTE_UNUSED marking.
5970 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5972 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5973 alternative and update.
5974 (aarch64_dup_lanedi): Delete.
5975 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5976 * config/aarch64/aarch64-simd-builtins.def: Update.
5978 2013-06-17 Richard Biener <rguenther@suse.de>
5980 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5981 (lto_input_scc): Declare.
5982 (lto_input_tree_1): Likewise.
5983 (struct lto_stats_d): Add num_tree_bodies_output and
5984 num_pickle_refs_output.
5985 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5986 (lto_read_tree_1): Split out from ...
5987 (lto_read_tree): ... this.
5988 (lto_input_scc): New function.
5989 (lto_input_tree_1): Split out from ...
5990 (lto_input_tree): ... this. Handle LTO_tree_scc.
5991 (lto_data_in_create): Create the streamer cache without hashes.
5992 * lto-streamer-out.c (create_output_block): Create the streamer
5993 cache with hashes when not doing WPA.
5994 (lto_write_tree_1): Split out from ...
5995 (lto_write_tree): ... this.
5996 (get_symbol_initial_value): New function.
5997 (lto_output_tree_1): Split out from ...
5998 (lto_output_tree): ... this. Write trees as series of SCCs
5999 using a DFS walk via DFS_write_tree.
6000 (struct sccs, struct scc_entry): New types.
6001 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6002 (DFS_write_tree_body): New function.
6003 (DFS_write_tree): Likewise.
6004 (hash_tree): Likewise.
6005 (scc_entry_compare): Likewise.
6006 (hash_scc): Likewise.
6007 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6008 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6009 TREE_CHAIN as regular reference.
6010 (streamer_read_integer_cst): Remove.
6011 (streamer_get_pickled_tree): Adjust.
6012 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6013 of DECL_EXTERNALs in BLOCK_VARS for now.
6014 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6016 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6017 Add hash value argument and record that if hashes are recorded
6019 (streamer_tree_cache_insert_1): Adjust.
6020 (streamer_tree_cache_insert): Likewise.
6021 (streamer_tree_cache_insert_at): Rename to ...
6022 (streamer_tree_cache_replace_tree): ... this and adjust.
6023 (streamer_tree_cache_append): Adjust.
6024 (record_common_node): Likewise.
6025 (streamer_tree_cache_create): Add argument whether to
6026 record hash values together with trees.
6027 (streamer_tree_cache_delete): Adjust.
6028 * tree-streamer.h (struct streamer_tree_cache_d): Add
6030 (streamer_read_integer_cst): Remove.
6031 (streamer_tree_cache_insert): Adjust.
6032 (streamer_tree_cache_append): Likewise.
6033 (streamer_tree_cache_insert_at): Rename to ...
6034 (streamer_tree_cache_replace_tree): ... this and adjust.
6035 (streamer_tree_cache_create): Add argument whether to record hashes.
6036 (streamer_tree_cache_get): Rename to ...
6037 (streamer_tree_cache_get_tree): ... this.
6038 (streamer_tree_cache_get_hash): New function.
6039 * tree.c (cache_integer_cst): New function.
6040 * tree.h (cache_integer_cst): Declare.
6041 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6042 * lto-symtab.c (lto_varpool_replace_node): Only release
6043 DECL_INITIAL of non-prevailing decls.
6044 * varpool.c (varpool_remove_initializer): Do not release
6045 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6047 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6049 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6050 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6051 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6052 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6053 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6054 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6055 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6056 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6057 instead of TARGET_64BIT.
6058 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6059 Require ISA_HAS_<D>DIV.
6061 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6063 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6064 (mips*-*-linux*): Move default with_llsc setting to where other
6066 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6068 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6069 Likewise. Remove default with_tune setting. Move default float
6070 setting to its own block. Handle with_llsc in the same block as above.
6072 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6074 PR rtl-optimization/57425
6075 PR rtl-optimization/57569
6076 * alias.c (write_dependence_p): Add new parameters mem_mode,
6077 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6078 Changed all callers.
6079 (canon_anti_dependence): New function.
6080 * cse.c (check_dependence): Use canon_anti_dependence.
6081 * cselib.c (cselib_invalidate_mem): Likewise.
6082 * rtl.h (canon_anti_dependence): Declare.
6084 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6086 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6087 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6088 ".set mips3" for 64-bit targets.
6090 2013-06-15 Dehao Chen <dehao@google.com>
6092 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6093 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6094 (gimple_check_call_args): Likewise.
6095 * value-prof.c (check_ic_target): Likewise.
6096 * ipa-inline.c (early_inliner): Likewise.
6097 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6098 * cgraph.c (cgraph_create_edge_1): Likewise.
6099 (cgraph_make_edge_direct): Likewise.
6101 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6104 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6105 rs6000_output_move_128bit to handle emitting quad memory
6106 operations. Set attribute length to 8 bytes.
6108 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6110 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6112 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6113 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6114 (aarch64_<su>mlsl<mode>): Likewise.
6116 2013-06-14 Mike Stump <mikestump@comcast.net>
6118 * Makefile.in (TARGET_H): Add insn-codes.h.
6120 2013-06-14 Alan Modra <amodra@gmail.com>
6124 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6125 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6126 bitfield expansion when EXPAND_MEMORY.
6127 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6129 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6131 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6132 test for clearing quad memory on 32-bit later.
6134 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6136 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6137 (fold_negate_expr): Likewise.
6138 (fold_real_zero_addition_p): Handle vectors.
6139 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6141 2013-06-14 Alan Modra <amodra@gmail.com>
6143 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6145 2013-06-13 Jan Hubicka <jh@suse.cz>
6147 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6148 Local comdats are not externally visible.
6149 * symtab.c (dump_symtab_base): Dump externally visible.
6150 (verify_symtab_base): Verify back links in the symtab hash.
6152 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6154 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6155 CONVERT_EXPR as equal nodes.
6157 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6159 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6161 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6163 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6164 Generalize to complex and vector.
6165 * tree.c (build_all_ones_cst): New function.
6166 * tree.h (build_all_ones_cst): Declare it.
6168 2013-06-13 Alan Modra <amodra@gmail.com>
6170 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6171 * config/rs6000/rs6000.md (signbittf2): New insn.
6172 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6173 (abstf2_internal, cmptf_internal2): Likewise.
6174 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6176 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6177 Pat Haugen <pthaugen@us.ibm.com>
6178 Peter Bergner <bergner@vnet.ibm.com>
6180 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6181 power8 byte, half-word, and quad-word atomic instructions.
6182 (emit_store_conditional): Likewise.
6183 (rs6000_expand_atomic_compare_and_swap): Likewise.
6184 (rs6000_expand_atomic_op): Likewise.
6186 * config/rs6000/sync.md (larx): Add new modes for power8.
6188 (AINT): New mode iterator to include TImode as well as normal
6189 integer modes on power8.
6190 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6191 that VSX registers are not considered. Use AINT mode iterator
6192 instead of INT1 to allow inclusion of quad word atomic operations
6194 (load_locked<mode>): Likewise.
6195 (store_conditional<mode>): Likewise.
6196 (atomic_compare_and_swap<mode>): Likewise.
6197 (atomic_exchange<mode>): Likewise.
6198 (atomic_nand<mode>): Likewise.
6199 (atomic_fetch_<fetchop_name><mode>): Likewise.
6200 (atomic_nand_fetch<mode>): Likewise.
6201 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
6203 (ATOMIC): On power8, add QImode, HImode modes.
6204 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
6205 modes that promote to SImode.
6206 (load_lockedti): Convert TImode arguments to PTImode, so that we
6207 get a guaranteed even/odd register pair.
6208 (load_lockedpti): Likewise.
6209 (store_conditionalti): Likewise.
6210 (store_conditionalpti): Likewise.
6212 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
6213 atomic load/store instructions.
6216 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
6218 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
6220 (insn_count): New attribute, with most cases extracted from...
6221 (length): ...here. Redefine most cases in terms of insn_count.
6222 (single_insn): Delete.
6223 (can_delay): Use insn_count to check for single instructions.
6224 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
6225 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
6226 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
6227 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
6228 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
6229 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
6230 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
6231 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
6232 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
6233 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
6234 rather than "length".
6235 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
6236 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
6237 Use "insn_count" rather than "length".
6238 * config/mips/mips-dsp.md
6239 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
6240 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
6243 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
6245 PR tree-optimization/57361
6246 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
6248 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
6250 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
6252 (aarch64_simd_combine<mode>): New instruction expansion.
6253 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
6255 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
6256 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
6258 2013-06-12 Jan Hubicka <jh@suse.cz>
6260 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
6261 decl has when in streaming stage.
6262 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
6263 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
6265 2013-06-12 Roland Stigge <stigge@antcom.de>
6268 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
6270 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6272 PR tree-optimization/57537
6273 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
6274 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
6276 2013-06-12 Richard Biener <rguenther@suse.de>
6278 * data-streamer.h (streamer_write_char_stream): CSE
6279 obs->current_pointer.
6280 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
6281 streamer_write_char_stream manually and optimize the resulting loop.
6282 (streamer_write_hwi_stream): Likewise.
6284 2013-06-12 Jan Hubicka <jh@suse.cz>
6286 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
6287 * cgraph.h (varpool_create_empty_node): Declare.
6288 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
6290 * symtab.c (symtab_unregister_node): Be lax about missin entries
6292 (symtab_get_node): Update comment.
6293 * varpool.c (varpool_create_empty_node): Break out from ...
6294 (varpool_node_for_decl): ... here.
6295 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
6297 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
6299 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
6300 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
6301 part. Use straight-line flow at the end.
6302 <COMPONENT_REF>: Remove superfluous else.
6303 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
6305 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6308 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
6309 target hook even for !TREE_PUBLIC decls. If no resolution info
6310 is available, return false for common and external decls.
6312 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6314 * config/rl78/constraints.md (U): New constraint.
6315 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
6318 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6321 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
6322 to allow returning address to AT_PLATFORM name.
6324 2013-06-11 Jan Hubicka <jh@suse.cz>
6326 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
6327 * cgraph.h (symtab_node_base): Add weakref flag.
6328 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
6329 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
6330 (output_weakrefs): Use weakref flag.
6331 * fold-const.c (simple_operand_p): Handle WEAK.
6332 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
6333 * ipa.c (varpool_externally_visible_p): Drop weakref.
6334 (function_and_variable_visibility): Update comment; fix weakref
6335 sanity checks; do not clear DECL_WEAK on them.
6336 * lto-cgraph.c (lto_output_node): update.
6337 (lto_output_varpool_node): Update.
6338 (input_overwrite_node): Update.
6339 (input_node): Update.
6340 (input_varpool_node): Update.
6341 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
6342 (lto_symtab_merge_symbols): Add sanity check.
6343 (lto_symtab_prevailing_decl): Do not special case weakrefs.
6344 * passes.c (rest_of_decl_compilation): Set static flag, too.
6345 * symtab.c (dump_symtab_base): Dump weakref.
6346 (verify_symtab_base): Sanity check weakrefs.
6347 (symtab_make_decl_local): Remove duplicated code.
6348 (symtab_alias_ultimate_target): Simplify.
6349 * varpool.c (varpool_create_variable_alias): Set weakref flag.
6351 2013-06-11 DJ Delorie <dj@redhat.com>
6353 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
6354 (rl78_unwind_word_mode): New.
6356 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6358 * final.c (debug_prefix_maps): Make static.
6360 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6362 * function.c (initial_trampoline): Remove stray copy.
6364 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
6366 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
6368 2013-06-11 Martin Jambor <mjambor@suse.cz>
6370 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
6371 within bounds at the beginning of the function.
6373 2013-06-11 Alan Modra <amodra@gmail.com>
6375 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
6377 (get_named_section): Don't NULL !DECL_P decl.
6379 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
6381 * doc/invoke.texi (core-avx2): Document.
6383 (atom): Updated with MOVBE.
6385 2013-06-11 Richard Biener <rguenther@suse.de>
6387 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
6389 2013-06-11 Anton Blanchard <anton@samba.org>
6391 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
6392 correct shift value in little-endian mode.
6394 2013-06-11 Jakub Jelinek <jakub@redhat.com>
6397 * varasm.c (get_variable_align): Move #endif to the right place.
6399 2013-06-10 Cary Coutant <ccoutant@google.com>
6401 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
6402 for hash so that hash table traversal order is deterministic.
6404 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
6405 Pat Haugen <pthaugen@us.ibm.com>
6406 Peter Bergner <bergner@vnet.ibm.com>
6408 * config/rs6000/vector.md (GPR move splitter): Do not split moves
6409 of vectors in GPRS if they are direct moves or quad word load or
6412 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
6414 (direct_move_p): Likewise.
6415 (quad_load_store_p): Likewise.
6417 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
6418 classes into bins based on the physical register type.
6419 (reg_class_to_reg_type): Likewise.
6420 (IS_STD_REG_TYPE): Likewise.
6421 (IS_FP_VECT_REG_TYPE): Likewise.
6422 (reload_fpr_gpr): Arrays to determine what insn to use if we can
6423 use direct move instructions.
6424 (reload_gpr_vsx): Likewise.
6425 (reload_vsx_gpr): Likewise.
6426 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
6427 information that is a simplification of register classes. Also
6428 precalculate direct move reload helpers.
6429 (direct_move_p): New function to return true if the operation can
6430 be done as a direct move instruciton.
6431 (quad_load_store_p): New function to return true if the operation
6432 is a quad memory operation.
6433 (rs6000_legitimize_address): If quad memory, only allow register
6434 indirect for TImode addresses.
6435 (rs6000_legitimate_address_p): Likewise.
6436 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
6437 (rs6000_reload_register_type): Likewise.
6438 (register_to_reg_type): Return register type.
6439 (rs6000_secondary_reload_simple_move): New helper function for
6440 secondary reload and secondary memory needed to identify anything
6441 that is a simple move, and does not need reloading.
6442 (rs6000_secondary_reload_direct_move): New helper function for
6443 secondary reload to identify cases that can be done with several
6444 instructions via the direct move instructions.
6445 (rs6000_secondary_reload_move): New helper function for secondary
6446 reload to identify moves between register types that can be done.
6447 (rs6000_secondary_reload): Add support for quad memory operations
6448 and for direct move.
6449 (rs6000_secondary_memory_needed): Likewise.
6450 (rs6000_debug_secondary_memory_needed): Change argument names.
6451 (rs6000_output_move_128bit): New function to return the move to
6452 use for 128-bit moves, including knowing about the various
6453 limitations of quad memory operations.
6455 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
6456 memory operations. call rs6000_output_move_128bit for the actual
6457 instruciton(s) to generate.
6458 (vsx_movti_64bit): Likewise.
6460 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
6461 (UNSPEC_P8V_MTVSRWZ): Likewise.
6462 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6463 (UNSPEC_P8V_MTVSRD): Likewise.
6464 (UNSPEC_P8V_XXPERMDI): Likewise.
6465 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6466 (UNSPEC_FUSION_GPR): Likewise.
6467 (FMOVE128_GPR): New iterator for direct move.
6468 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6471 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6472 loads and direct move instructions.
6473 (zero_extendsidi2_lfiwzx): Likewise.
6474 (extendsidi2_lfiwax): Likewise.
6475 (extendsidi2_nocell): Likewise.
6476 (floatsi<mode>2_lfiwax): Likewise.
6478 (floatunssi<mode>2_lfiwzx): Likewise.
6480 (fix_trunc<mode>_stfiwx): Likewise.
6481 (fixuns_trunc<mode>_stfiwx): Likewise.
6482 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6483 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6484 (parity<mode>2_cmpb): Set length/type attr.
6485 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6486 for 'mr.' to fast_compare.
6487 (bpermd_<mode>): Change type attr to popcnt.
6488 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6489 (p8_mtvsrwz_1): Likewise.
6490 (p8_mtvsrwz_2): Likewise.
6491 (reload_fpr_from_gpr<mode>): Likewise.
6492 (p8_mtvsrd_1): Likewise.
6493 (p8_mtvsrd_2): Likewise.
6494 (p8_xxpermdi_<mode>): Likewise.
6495 (reload_vsx_from_gpr<mode>): Likewise.
6496 (reload_vsx_from_gprsf): Likewise.
6497 (p8_mfvsrd_3_<mode>): LIkewise.
6498 (reload_gpr_from_vsx<mode>): Likewise.
6499 (reload_gpr_from_vsxsf): Likewise.
6500 (p8_mfvsrd_4_disf): Likewise.
6501 (multi-word GPR splits): Do not split direct moves or quad memory
6504 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6506 * tree-into-ssa.c (interesting_blocks): Make static.
6508 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6511 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6512 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6513 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6514 (get_variable_align): New function.
6515 (get_variable_section, emit_bss, emit_common,
6516 assemble_variable_contents, place_block_symbol): Use
6517 get_variable_align instead of DECL_ALIGN.
6518 (assemble_noswitch_variable): Add align argument, use it
6519 instead of DECL_ALIGN.
6520 (assemble_variable): Adjust caller. Use get_variable_align
6521 instead of DECL_ALIGN.
6522 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6524 (DATA_ABI_ALIGNMENT): Define.
6525 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6526 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6527 opt is false, only return the psABI mandated alignment increase.
6528 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6529 (DATA_ABI_ALIGNMENT): ... this.
6530 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6531 (DATA_ABI_ALIGNMENT): ... this.
6532 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6533 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6534 (DATA_ABI_ALIGNMENT): ... this.
6535 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6536 * doc/tm.texi: Regenerated.
6538 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6540 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6541 cmp_code to construct REG_EQUAL note.
6543 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6546 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6547 that operands[2] doesn't overlap with operands[0].
6549 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6550 Jan Hubicka <jh@suse.cz>
6552 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6553 hack to mark symbols as used.
6555 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6557 PR rtl-optimization/57559
6558 * lra-constraints.c (process_alt_operands): Don't discourage
6559 memory with known offset for offsetable memory constraint.
6560 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6562 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6564 * varasm.c (struct oc_local_state): Reorder fields.
6565 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6566 and adjust accordingly.
6567 (output_constructor): Reorder initialization code and adjust call to
6568 output_constructor_bitfield.
6570 2013-06-07 Jan Hubicka <jh@suse.cz>
6572 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6574 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6576 * tree-object-size.c (unknown): Make const.
6578 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6580 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6581 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6582 for last alternative in the cpu_facility attribute.
6584 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6587 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6588 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6589 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6590 * config/arm/constraints.md (Dg): New constraint.
6591 * config/arm/neon.md (xordi3_neon): Remove.
6592 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6593 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6595 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6597 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6598 Clean up alternatives.
6600 2013-06-07 Alan Modra <amodra@gmail.com>
6602 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6605 2013-06-07 Alan Modra <amodra@gmail.com>
6607 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6609 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6611 * config/arm/constraints.md (Df): New constraint.
6612 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6613 Correct length attribute for last two alternatives.
6615 2013-06-07 Alan Modra <amodra@gmail.com>
6617 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6618 override user -mfp-in-toc.
6619 (offsettable_ok_by_alignment): Consider just the current access
6620 rather than the whole object, unless BLKmode. Handle
6621 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6622 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6623 for -mcmodel=medium.
6624 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6625 override user -mfp-in-toc or -msum-in-toc. Default to
6626 -mno-fp-in-toc for -mcmodel=medium.
6628 2013-06-06 DJ Delorie <dj@redhat.com>
6630 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6631 TARGET_VALID_POINTER_MODE.
6633 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6634 Pat Haugen <pthaugen@us.ibm.com>
6635 Peter Bergner <bergner@vnet.ibm.com>
6637 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6638 Document new power8 builtins.
6640 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6641 condition code register, to allow 128-bit logical operations to be
6642 done in the VSX or GPR registers.
6643 (nor<mode>3): Use the canonical form for nor.
6644 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6645 vclz*, and vpopcnt* vector instructions.
6646 (nand<mode>3): Likewise.
6647 (orc<mode>3): Likewise.
6648 (clz<mode>2): LIkewise.
6649 (popcount<mode>2): Likewise.
6651 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6652 that only the GPRs are recognized.
6654 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6655 support for new power8 builtins.
6657 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6659 (xscvdpspn): Likewise.
6665 (vpopcnt): Likewise.
6666 (vpopcntb): Likewise.
6667 (vpopcnth): Likewise.
6668 (vpopcntw): Likewise.
6669 (vpopcntd): Likewise.
6674 (eqv_v16qi3): Likewise.
6675 (eqv_v8hi3): Likewise.
6676 (eqv_v4si3): Likewise.
6677 (eqv_v2di3): Likewise.
6678 (eqv_v4sf3): Likewise.
6679 (eqv_v2df3): Likewise.
6681 (nand_v16qi3): Likewise.
6682 (nand_v8hi3): Likewise.
6683 (nand_v4si3): Likewise.
6684 (nand_v2di3): Likewise.
6685 (nand_v4sf3): Likewise.
6686 (nand_v2df3): Likewise.
6688 (orc_v16qi3): Likewise.
6689 (orc_v8hi3): Likewise.
6690 (orc_v4si3): Likewise.
6691 (orc_v2di3): Likewise.
6692 (orc_v4sf3): Likewise.
6693 (orc_v2df3): Likewise.
6695 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6696 allow power8 quad mode in 64-bit.
6697 (rs6000_builtin_vectorized_function): Add support to vectorize
6698 ISA 2.07 count leading zeros, population count builtins.
6699 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6700 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6701 (builtin_function_type): Add vgbbd builtin function which takes an
6703 (altivec_expand_vec_perm_const): Add support for new power8 merge
6706 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6707 that does not include TImdoe for use with 32-bit.
6708 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6710 (UNSPEC_VSX_CVDPSPN): Likewise.
6711 (vsx_xscvdpspn): Likewise.
6712 (vsx_xscvspdpn): Likewise.
6713 (vsx_xscvdpspn_scalar): Likewise.
6714 (vsx_xscvspdpn_directmove): Likewise.
6715 (vsx_and<mode>3): Split logical operations into 32-bit and
6716 64-bit. Add support to do logical operations on TImode as well as
6717 VSX vector types. Allow logical operations to be done in either
6718 VSX registers or in general purpose registers in 64-bit mode. Add
6719 splitters if GPRs were used. For AND, add clobber of CCmode to
6720 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6722 (vsx_and<mode>3_32bit): Likewise.
6723 (vsx_and<mode>3_64bit): Likewise.
6724 (vsx_ior<mode>3): Likewise.
6725 (vsx_ior<mode>3_32bit): Likewise.
6726 (vsx_ior<mode>3_64bit): Likewise.
6727 (vsx_xor<mode>3): Likewise.
6728 (vsx_xor<mode>3_32bit): Likewise.
6729 (vsx_xor<mode>3_64bit): Likewise.
6730 (vsx_one_cmpl<mode>2): Likewise.
6731 (vsx_one_cmpl<mode>2_32bit): Likewise.
6732 (vsx_one_cmpl<mode>2_64bit): Likewise.
6733 (vsx_nor<mode>3): Likewise.
6734 (vsx_nor<mode>3_32bit): Likewise.
6735 (vsx_nor<mode>3_64bit): Likewise.
6736 (vsx_andc<mode>3): Likewise.
6737 (vsx_andc<mode>3_32bit): Likewise.
6738 (vsx_andc<mode>3_64bit): Likewise.
6739 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6740 and xxlorc instructions.
6741 (vsx_eqv<mode>3_64bit): Likewise.
6742 (vsx_nand<mode>3_32bit): Likewise.
6743 (vsx_nand<mode>3_64bit): Likewise.
6744 (vsx_orc<mode>3_32bit): Likewise.
6745 (vsx_orc<mode>3_64bit): Likewise.
6747 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6749 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6751 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6752 (p8_vmrgow): Likewise.
6753 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6754 GPRs to be split under VSX.
6755 (p8v_clz<mode>2): Add power8 count leading zero support.
6756 (p8v_popcount<mode>2): Add power8 population count support.
6757 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6760 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6763 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6765 (vec_nand): Likewise.
6766 (vec_vclz): Likewise.
6767 (vec_vclzb): Likewise.
6768 (vec_vclzd): Likewise.
6769 (vec_vclzh): Likewise.
6770 (vec_vclzw): Likewise.
6771 (vec_vgbbd): Likewise.
6772 (vec_vmrgew): Likewise.
6773 (vec_vmrgow): Likewise.
6774 (vec_vpopcnt): Likewise.
6775 (vec_vpopcntb): Likewise.
6776 (vec_vpopcntd): Likewise.
6777 (vec_vpopcnth): Likewise.
6778 (vec_vpopcntw): Likewise.
6780 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6782 PR rtl-optimization/57468
6783 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6786 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6788 PR rtl-optimization/57459
6789 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6790 type when setting live regs.
6792 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6794 * config/s390/s390.opt (mlra): New option.
6795 * config/s390/s390.c (s390_decompose_address): Check displacement
6796 for all registers for LRA.
6797 (s390_secondary_reload): Don't used secondary reloads for LRA.
6798 (s390_lra_p): New function.
6799 (TARGET_LRA_P): Define.
6800 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6801 of attribute cpu_facility to zarch for the last alternative.
6802 (*cmpmem_short): Ditto.
6804 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6806 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6807 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6808 (arm_expand_prologue): Likewise.
6810 2013-06-06 Teresa Johnson <tejohnson@google.com>
6813 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6814 as this is now done by redirect_edge_and_branch_force.
6815 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6816 barriers, and fix interaction with splitting.
6817 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6818 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6819 reflect changes made in the routine.
6820 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6821 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6822 since this is called in cfglayout mode, and replace partition fixup
6823 with assert as that is now done by force_nonfallthru_and_redirect.
6824 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6825 already be marked with region crossing note.
6826 (insert_section_boundary_note): Make non-static, gate on flag
6827 has_bb_partition, rewrite to also check for multiple partitions.
6828 (rest_of_handle_reorder_blocks): Remove call to
6829 insert_section_boundary_note, now done later during free_cfg.
6830 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6831 * bb-reorder.h (insert_section_boundary_note): Declare.
6832 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6833 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6834 invoke insert_section_boundary_note.
6835 (try_redirect_by_replacing_jump): Remove unnecessary
6836 check for region crossing note.
6837 (fixup_partition_crossing): New function.
6838 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6839 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6840 in non-cfglayout mode.
6841 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6842 remove old code that tried to do this. Emit barrier correctly
6843 when we are in cfglayout mode.
6844 (last_bb_in_partition): New function.
6845 (rtl_split_edge): Correctly fixup partition boundaries.
6846 (commit_one_edge_insertion): Remove old code that tried to
6847 fixup region crossing edge since this is now handled in
6848 split_block, and set up insertion point correctly since
6849 block may now end in a jump.
6850 (verify_hot_cold_block_grouping): Guard against checking when not in
6851 linearized RTL mode.
6852 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6854 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6855 rtl_verify_flow_info, so not called in cfglayout mode.
6856 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6857 (fixup_reorder_chain): Remove old code that attempted to fixup region
6858 crossing note as this is now handled in force_nonfallthru_and_redirect.
6859 (duplicate_insn_chain): Don't duplicate switch section notes.
6860 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6862 * basic-block.h (emit_barrier_after_bb): Declare.
6864 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6866 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6867 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6868 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6870 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6872 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6874 * config/arm/ldmstm.md: Regenerate.
6876 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6878 * config/arm/sync.md (atomic_loaddi_1):
6879 Disable predication for arm_restrict_it.
6880 (arm_load_exclusive<mode>): Likewise.
6881 (arm_load_exclusivesi): Likewise.
6882 (arm_load_exclusivedi): Likewise.
6883 (arm_load_acquire_exclusive<mode>): Likewise.
6884 (arm_load_acquire_exclusivesi): Likewise.
6885 (arm_load_acquire_exclusivedi): Likewise.
6886 (arm_store_exclusive<mode>): Likewise.
6887 (arm_store_exclusive<mode>): Likewise.
6888 (arm_store_release_exclusivedi): Likewise.
6889 (arm_store_release_exclusive<mode>): Likewise.
6891 2013-06-06 Richard Biener <rguenther@suse.de>
6893 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6895 (lto_tag_is_tree_code_p): Adjust.
6896 (lto_tag_is_gimple_code_p): Likewise.
6897 (lto_gimple_code_to_tag): Likewise.
6898 (lto_tag_to_gimple_code): Likewise.
6899 (lto_tree_code_to_tag): Likewise.
6900 (lto_tag_to_tree_code): Likewise.
6901 * data-streamer.h (streamer_write_hwi_in_range): Use
6902 uhwi streaming to stream the normalized range.
6903 (streamer_read_hwi_in_range): Likewise.
6905 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6907 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6908 (predicable_short_it): Likewise.
6909 (predicated): Likewise.
6910 (enabled): Handle above.
6911 (define_cond_exec): Set predicated attribute to yes.
6913 2013-06-05 Mike Stump <mikestump@comcast.net>
6915 * gdbinit.in (__FUNCTION__): Add.
6917 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6919 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6920 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6922 2013-06-05 Jan Hubicka <jh@suse.cz>
6924 * varasm.c (mark_decl_referenced): Revert the removal until targets
6927 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6929 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6930 instead of mark_decl_referenced.
6932 2013-06-05 Jan Hubicka <jh@suse.cz>
6934 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6935 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6936 and symtab_used_from_object_file_p.
6937 (cgraph_make_node_local_1): Clear forced_by_abi.
6938 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6939 * cgraph.h (symtab_node_base): Add forced_by_abi.
6940 (decide_is_variable_needed): Remove.
6941 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6942 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6943 (decide_is_symbol_needed): ... this one; handle symbols in general;
6944 always analyze virtuals; honnor forced_by_abi.
6945 (cgraph_finalize_function): Update.
6946 (varpool_finalize_decl): Update.
6947 (symbol_defined_and_needed): Remove.
6948 (analyze_functions): Update.
6949 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6950 output_refs, input_overwrite_node): Handle forced_by_abi.
6951 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6952 (address_taken_from_non_vtable_p): ... this one.
6953 (comdat_can_be_unshared_p_1): New function.
6954 (cgraph_comdat_can_be_unshared_p): Rename to ...
6955 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6956 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6957 (function_and_variable_visibility): Clear forced_by_abi as needed.
6958 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6959 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6960 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6961 * varpool.c (decide_is_variable_needed): Remove.
6963 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6965 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6966 (arm_option_override): Override arm_restrict_it where appropriate.
6967 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6968 * config/arm/arm.opt (mrestrict-it): New command-line option.
6969 * doc/invoke.texi: Document -mrestrict-it.
6971 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6973 * tsan.c (tsan_atomic_table): Make const.
6975 2013-06-05 Richard Biener <rguenther@suse.de>
6977 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6978 index associated with the tree we are supposed to replace.
6979 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6980 TREE_ASM_WRITTEN as zero for everything but SSA names.
6982 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6984 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6986 2013-06-04 Jan Hubicka <jh@suse.cz>
6988 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6989 (input_node, input_varpool_node): Handle correctly external same
6991 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6992 nodes at ltrans stage.
6994 2013-06-04 Jan Hubicka <jh@suse.cz>
6996 * ipa-inline.c (update_caller_keys): Fix availability test.
6997 (update_callee_keys): Likewise.
6998 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6999 to follow ELF standard.
7001 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7003 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7004 (mips64r5900el-*-elf*): New configurations.
7005 * config/mips/mips-cpus.def (r5900): New processor.
7006 * config/mips/mips-tables.opt: Regenerate.
7007 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7008 (mips_issue_rate): Handle PROCESSOR_R5900.
7009 (mips_reorg_process_insns): Force reorder mode for the R5900.
7010 * config/mips/mips.h (TARGET_MIPS5900): Define.
7011 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7013 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7015 * config/mips/mips.md (processor): Add r5900.
7016 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7018 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7020 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7021 into function to generate MOVI instruction.
7022 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7023 (aarch64_preferred_simd_mode): Turn into wrapper.
7024 (aarch64_output_scalar_simd_mov_immediate): New function.
7025 * config/aarch64/aarch64-protos.h: Add prototype for above.
7027 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7029 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7030 element_char member.
7031 (sizetochar): Return signed char.
7032 (aarch64_simd_valid_immediate): Remove elchar and other
7033 unnecessary variables.
7034 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7035 Calculate element_char as required.
7036 * config/aarch64/aarch64-protos.h: Update and move prototype
7037 for aarch64_output_simd_mov_immediate.
7038 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7041 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7043 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7044 information completed by aarch64_simd_valid_immediate.
7045 (aarch64_legitimate_constant_p): Update arguments.
7046 (aarch64_simd_valid_immediate): Work with struct rather than many
7048 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7049 (aarch64_simd_make_constant): Update arguments.
7050 (aarch64_output_simd_mov_immediate): Work with struct rather than
7051 many pointers. Output immediate directly rather than as operand.
7052 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7054 * config/aarch64/constraints.md (Dn): Update arguments.
7056 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7058 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7060 (aarch64_simd_immediate_valid_for_move): Remove.
7061 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7062 (aarch64_simd_make_constant): Update call.
7063 (aarch64_output_simd_mov_immediate): Update call.
7064 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7066 * config/aarch64/constraints.md (Dn): Update call.
7068 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7070 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7071 return type to bool for prototype.
7072 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7073 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7074 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7076 2013-06-04 Catherine Moore <clm@codesourcery.com>
7078 * config/mips/mips.opt (meva): New.
7079 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7080 (ASM_SPEC): Handle -meva.
7081 * doc/invoke.texi (meva): Document.
7083 2013-06-04 Alan Modra <amodra@gmail.com>
7085 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7088 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7090 * rtl.def: Add extra fourth optional field to define_cond_exec.
7091 * gensupport.c (process_one_cond_exec): Process attributes from
7093 * doc/md.texi: Document fourth field in define_cond_exec.
7095 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7097 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7098 out the processing order as in store_bit_field_1.
7100 2013-06-04 Jan Hubicka <jh@suse.cz>
7103 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7104 non-VAR_DECL node if it does not exist yet.
7106 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7108 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7109 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7110 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7111 target_cpu_default setting.
7113 2013-06-03 Teresa Johnson <tejohnson@google.com>
7115 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7116 default to -fopt-info=optimized instead of all.
7117 * doc/invoke.texi: Ditto.
7118 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7119 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7120 (execute_vect_slp): Emit BB vectorization success under
7121 MSG_OPTIMIZED_LOCATIONS.
7122 * tree-vect-slp.c (vect_slp_transform_bb): Change
7123 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7124 * tree-vect-loop.c (vect_transform_loop): Ditto.
7126 2013-06-03 Jason Merrill <jason@redhat.com>
7129 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7130 Use TARGET_EXPR for C++.
7132 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7134 PR rtl-optimization/57268
7135 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7136 if DEBUG_INSN_P (insn).
7139 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7141 PR rtl-optimization/57268
7142 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7143 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7145 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7147 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7148 (ix86_avoid_lea_for_addr): Likewise.
7149 (exact_dependency_1): Likewise.
7150 (ix86_adjust_cost): Likewise.
7151 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7153 (do_reorder_for_imul): Fix typo, formatting and
7154 !reload_completed check removed.
7155 (ix86_sched_reorder): Fix typo and formatting.
7156 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7159 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7161 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7163 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7165 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7166 <STRING_CST>: Likewise.
7167 <VECTOR_CST>: Likewise.
7169 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7170 Mikael Morin <mikael@gcc.gnu.org>
7172 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7173 * config.in: Regenerated.
7174 * configure: Regenerated.
7176 2013-06-01 Jan Hubicka <jh@suse.cz>
7179 * cgraphunit.c (compile): When weakref is not supported,
7180 set up transparent aliases before final output pass.
7181 * varasm.c (assemble_alias): Do not try to do it here.
7183 2013-06-01 Jan Hubicka <jh@suse.cz>
7186 * passes.c (for_per_function): Skip unanalyzed functions.
7188 2013-06-01 Jan Hubicka <jh@suse.cz>
7190 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7191 (lto_symtab_merge_symbols_1): ... this one.
7192 (lto_symtab_merge_cgraph_nodes): Rename to ...
7193 (lto_symtab_merge_symbols): ... this one; simplify.
7194 * cgraph.c (same_body_aliases_done): Rename to ...
7195 (cpp_implicit_aliases_done): ... this one.
7196 (cgraph_create_function_alias): Update.
7197 (cgraph_same_body_alias): Update.
7198 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
7199 (verify_edge_corresponds_to_fndecl): Simplify.
7200 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
7201 (cgraph_node): Remove same_body_alias.
7202 (varpool_node): Remove alias_of and extra_name_alias.
7203 (same_body_aliases_done): Rename to ..
7204 (cpp_implicit_aliases_done): ... this one.
7205 (symtab_alias_ultimate_target): Add default parameter.
7206 (symtab_resolve_alias): New function.
7207 (fixup_same_cpp_alias_visibility): Declare.
7208 (cgraph_function_node): Add default parameter.
7209 (cgraph_node_asm_name): Likewise.
7210 (cgraph_function_or_thunk_node): Add default parameter; do
7211 not ICE when it is NULL.
7212 (varpool_variable_node): Likewise.
7213 * tree-emutls.c (create_emultls_var): Update.
7214 (ipa_lower_emutls): Update.
7215 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
7216 (cgraph_reset_node): Reset alias info.
7217 (cgraph_finalize_function): Update.
7218 (fixup_same_cpp_alias_visibility): Move to symtab.c.
7219 (analyze_function): Simplify.
7220 (cgraph_process_same_body_aliases): Simplify.
7221 (analyze_functions): Fixup same body aliases.
7222 (handle_alias_pairs): Simplify.
7223 (assemble_thunk): Update.
7224 (assemble_thunks_and_aliases): Update.
7225 (output_weakrefs): Rewrite.
7226 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
7227 (lto_output_varpool_node): Likewise.
7228 (compute_ltrans_boundary): Remve assert.
7229 (get_alias_symbol): New functoin.
7230 (input_node): Rewrite alias handling.
7231 (input_varpool_node): Likewise.
7232 * ipa-pure-const.c (propagate_pure_const): Fix formating.
7233 * ipa.c (process_references): Handle weakrefs correctly.
7234 (symtab_remove_unreachable_nodes): Likewise.
7235 * trans-mem.c (get_cg_data): Update.
7236 (ipa_tm_create_version_alias): Update.
7237 (ipa_tm_execute): Update.
7238 * symtab.c (dump_symtab_base): Dump aliases.
7239 (verify_symtab_base): Verify aliases.
7240 (symtab_node_availability): New function.
7241 (symtab_alias_ultimate_target): Simplify.
7242 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
7243 handle all the fixup cases.
7244 (symtab_resolve_alias): New function.
7245 * passes.c (ipa_write_summaries): Handle weakrefs.
7246 * varpool.c (varpool_analyze_node): Simplify.
7247 (assemble_aliases): Update.
7248 (varpool_create_variable_alias): Simplify.
7249 (varpool_extra_name_alias): Simplify.
7250 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
7251 (lto_symtab_merge_symbols): ... this one.
7253 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
7256 PR rtl-optimization/57268
7257 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7258 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7260 2013-06-01 Tobias Burnus <burnus@net-b.de>
7263 2013-05-31 Tobias Burnus <burnus@net-b.de>
7266 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
7269 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7271 PR rtl-optimization/57268
7272 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7273 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7275 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7277 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
7278 unordered comparison operators when -fno-trapping-math is in effect
7280 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
7281 and implement unordered comparison operators properly on the e500.
7283 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7285 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
7286 for constant scalar integers.
7287 (simplify_relational_operation_1): Likewise.
7289 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
7291 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
7292 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
7295 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7296 Igor Zamyatin <igor.zamyatin@intel.com>
7298 Silvermont (SLM) architecture performance tuning.
7299 * config/i386/i386.h (enum ix86_tune_indices): Add
7300 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
7301 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
7303 * config/i386/i386.c (initial_ix86_tune_features)
7304 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
7305 (ix86_lea_outperforms): Handle Silvermont tuning.
7306 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
7308 (ix86_use_lea_for_mov): Likewise.
7309 (ix86_avoid_lea_for_addr): Likewise.
7310 (ix86_lea_for_add_ok): Likewise.
7311 (exact_dependency_1): New function.
7312 (exact_store_load_dependency): Likewise.
7313 (ix86_adjust_cost): Handle Silvermont tuning.
7314 (do_reoder_for_imul): Likewise.
7315 (swap_top_of_ready_list): New function.
7316 (ix86_sched_reorder): Changed to handle Silvermont tuning.
7318 * config/i386/i386.md (peepholes that split memory operand in fp
7321 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7323 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
7324 Remove un-necessary braces.
7326 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7328 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7329 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
7331 2013-05-31 Tobias Burnus <burnus@net-b.de>
7334 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
7336 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7339 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
7340 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
7341 * config/arm/neon.md (iordi3_neon): Remove.
7342 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
7343 * config/arm/predicates.md (imm_for_neon_logic_operand):
7344 Move to earlier in the file.
7345 (neon_logic_op2): Likewise.
7346 (arm_iordi_operand_neon): New predicate.
7348 2013-05-31 Richard Biener <rguenther@suse.de>
7350 PR tree-optimization/57478
7351 PR tree-optimization/57453
7352 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
7355 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7357 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
7358 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
7360 2013-05-30 Tobias Burnus <burnus@net-b.de>
7361 Thomas Koenig <tkoenig@gcc.gnu.org>
7364 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
7365 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
7367 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
7369 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
7371 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
7373 * target.def (register_usage_leveling_p): New hook.
7374 * targhooks.c (default_register_usage_leveling_p): New.
7375 * targhooks.h (default_register_usage_leveling_p): New prototype.
7376 * lra-assigns.c (register_usage_leveling_p): Use the hook.
7377 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
7378 * doc/tm.texi: Update.
7379 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
7381 2013-05-30 Ian Bolton <ian.bolton@arm.com>
7383 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
7384 (*insv_reg<mode>): New define_insn.
7386 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
7388 PR rtl-optimization/57439
7389 * postreload.c (move2add_valid_value_p): Check that we have
7390 a zero subreg_regno_offset when accessing the register in
7393 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7394 Igor Zamyatin <igor.zamyatin@intel.com>
7396 Silvermont (SLM) architecture pipeline model, tuning and
7398 * config.gcc: Add slm config options and target.
7400 * config/i386/slm.md: New.
7402 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
7404 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
7406 (ix86_target_macros_internal): Likewise.
7408 * gcc/config/i386/i386.c (slm_cost): New cost.
7409 (m_SLM): New macro flag.
7410 (initial_ix86_tune_features): Set m_SLM.
7411 (x86_accumulate_outgoing_args): Likewise.
7412 (x86_arch_always_fancy_math_387): Likewise.
7413 (processor_target_table): Add slm cost.
7414 (cpu_names): Add slm cpu name.
7415 (x86_option_override_internal): Set SLM ISA.
7416 (ix86_issue_rate): New case PROCESSOR_SLM.
7417 (ia32_multipass_dfa_lookahead): Likewise.
7418 (fold_builtin_cpu): Add slm.
7420 * config/i386/i386.h (TARGET_SLM): New target macro.
7421 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
7422 (processor_type): Add PROCESSOR_SLM.
7424 * config/i386/i386.md (cpu): Add new value "slm".
7425 (slm.md): Include slm.md.
7427 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
7428 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7430 * config/arm/arm-protos.h: Add and update function protos.
7431 * config/arm/arm.c (use_simple_return_p): New added.
7432 (thumb2_expand_return): Check simple_return flag.
7433 * config/arm/arm.md: Add simple_return and conditional simple_return.
7434 * config/arm/iterators.md: Add iterator for return and simple_return.
7436 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7438 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
7439 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
7440 (arm_emit_vfp_multi_reg_pop): Likewise.
7441 (thumb2_emit_ldrd_pop): Likewise.
7442 (arm_expand_epilogue): Add misc REG_CFA notes.
7443 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
7445 2013-05-29 Lawrence Crowl <crowl@google.com>
7447 * config/arm/t-arm: Update for below.
7449 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
7450 Change type to hash_table. Update dependent calls and types.
7452 * config/i386/t-cygming: Update for below.
7454 * config/i386/t-interix: Update for below.
7456 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
7457 Change type to hash_table. Update dependent calls and types.
7458 (i386_find_on_wrapper_list::wrappers): Likewise.
7460 * config/ia64/t-ia64: Update for below.
7462 * config/ia64/ia64.c (bundle_state_table):
7463 Change type to hash_table. Update dependent calls and types.
7465 * config/mips/mips.c (mips_reorg_process_insns::htab):
7466 Change type to hash_table. Update dependent calls and types.
7468 * config/sol2.c (solaris_comdat_htab):
7469 Change type to hash_table. Update dependent calls and types.
7471 * config/t-sol2: Update for above.
7473 2013-05-29 Teresa Johnson <tejohnson@google.com>
7475 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7476 functions are not yet marked as defined.
7478 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7479 Pat Haugen <pthaugen@us.ibm.com>
7480 Peter Bergner <bergner@vnet.ibm.com>
7482 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7486 (vrotl<mode>3): Likewise.
7487 (vashl<mode>3): Likewise.
7488 (vlshr<mode>3): Likewise.
7489 (vashr<mode>3): Likewise.
7491 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7492 support for power8 V2DI builtins.
7494 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7495 power8 V2DI builtins.
7496 (vupkhsw): Likewise.
7497 (vupklsw): Likewise.
7498 (vaddudm): Likewise.
7503 (vpkudum): Likewise.
7504 (vpksdss): Likewise.
7505 (vpkudus): Likewise.
7506 (vpksdus): Likewise.
7511 (vsubudm): Likewise.
7512 (vcmpequd): Likewise.
7513 (vcmpgtsd): Likewise.
7514 (vcmpgtud): Likewise.
7515 (vcmpequd_p): Likewise.
7516 (vcmpgtsd_p): Likewise.
7517 (vcmpgtud_p): Likewise.
7518 (vupkhsw): Likewise.
7519 (vupklsw): Likewise.
7520 (vaddudm): Likewise.
7525 (vpksdss): Likewise.
7526 (vpksdus): Likewise.
7527 (vpkudum): Likewise.
7528 (vpkudus): Likewise.
7533 (vsubudm): Likewise.
7535 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7536 support for power8 V2DI instructions.
7538 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7539 power8 V2DI instructions. Combine pack and unpack insns to use an
7540 iterator for each mode. Check whether a particular mode supports
7541 Altivec instructions instead of just checking TARGET_ALTIVEC.
7542 (UNSPEC_VPKUWUM): Likewise.
7543 (UNSPEC_VPKSHSS): Likewise.
7544 (UNSPEC_VPKSWSS): Likewise.
7545 (UNSPEC_VPKUHUS): Likewise.
7546 (UNSPEC_VPKSHUS): Likewise.
7547 (UNSPEC_VPKUWUS): Likewise.
7548 (UNSPEC_VPKSWUS): Likewise.
7549 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7550 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7551 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7552 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7553 (UNSPEC_VUPKHSB): Likewise.
7554 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7555 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7556 (UNSPEC_VUPKHSH): Likewise.
7557 (UNSPEC_VUPKLSB): Likewise.
7558 (UNSPEC_VUPKLSH): Likewise.
7560 (VI_char): Likewise.
7561 (VI_scalar): Likewise.
7562 (VI_unit): Likewise.
7564 (VP_small): Likewise.
7565 (VP_small_lc): Likewise.
7566 (VU_char): Likewise.
7567 (add<mode>3): Likewise.
7568 (altivec_vaddcuw): Likewise.
7569 (altivec_vaddu<VI_char>s): Likewise.
7570 (altivec_vadds<VI_char>s): Likewise.
7571 (sub<mode>3): Likewise.
7572 (altivec_vsubcuw): Likewise.
7573 (altivec_vsubu<VI_char>s): Likewise.
7574 (altivec_vsubs<VI_char>s): Likewise.
7575 (altivec_vavgs<VI_char>): Likewise.
7576 (altivec_vcmpbfp): Likewise.
7577 (altivec_eq<mode>): Likewise.
7578 (altivec_gt<mode>): Likewise.
7579 (altivec_gtu<mode>): Likewise.
7580 (umax<mode>3): Likewise.
7581 (smax<mode>3): Likewise.
7582 (umin<mode>3): Likewise.
7583 (smin<mode>3): Likewise.
7584 (altivec_vpkuhum): Likewise.
7585 (altivec_vpkuwum): Likewise.
7586 (altivec_vpkshss): Likewise.
7587 (altivec_vpkswss): Likewise.
7588 (altivec_vpkuhus): Likewise.
7589 (altivec_vpkshus): Likewise.
7590 (altivec_vpkuwus): Likewise.
7591 (altivec_vpkswus): Likewise.
7592 (altivec_vpks<VI_char>ss): Likewise.
7593 (altivec_vpks<VI_char>us): Likewise.
7594 (altivec_vpku<VI_char>us): Likewise.
7595 (altivec_vpku<VI_char>um): Likewise.
7596 (altivec_vrl<VI_char>): Likewise.
7597 (altivec_vsl<VI_char>): Likewise.
7598 (altivec_vsr<VI_char>): Likewise.
7599 (altivec_vsra<VI_char>): Likewise.
7600 (altivec_vsldoi_<mode>): Likewise.
7601 (altivec_vupkhsb): Likewise.
7602 (altivec_vupkhs<VU_char>): Likewise.
7603 (altivec_vupkls<VU_char>): Likewise.
7604 (altivec_vupkhsh): Likewise.
7605 (altivec_vupklsb): Likewise.
7606 (altivec_vupklsh): Likewise.
7607 (altivec_vcmpequ<VI_char>_p): Likewise.
7608 (altivec_vcmpgts<VI_char>_p): Likewise.
7609 (altivec_vcmpgtu<VI_char>_p): Likewise.
7610 (abs<mode>2): Likewise.
7611 (vec_unpacks_hi_v16qi): Likewise.
7612 (vec_unpacks_hi_v8hi): Likewise.
7613 (vec_unpacks_lo_v16qi): Likewise.
7614 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7615 (vec_unpacks_lo_v8hi): Likewise.
7616 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7617 (vec_pack_trunc_v8h): Likewise.
7618 (vec_pack_trunc_v4si): Likewise.
7619 (vec_pack_trunc_<mode>): Likewise.
7621 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7623 (vec_vmaxsd): Likewise.
7624 (vec_vmaxud): Likewise.
7625 (vec_vminsd): Likewise.
7626 (vec_vminud): Likewise.
7627 (vec_vpksdss): Likewise.
7628 (vec_vpksdus): Likewise.
7629 (vec_vpkudum): Likewise.
7630 (vec_vpkudus): Likewise.
7631 (vec_vrld): Likewise.
7632 (vec_vsld): Likewise.
7633 (vec_vsrad): Likewise.
7634 (vec_vsrd): Likewise.
7635 (vec_vsubudm): Likewise.
7636 (vec_vupkhsw): Likewise.
7637 (vec_vupklsw): Likewise.
7639 2013-05-29 Jan Hubicka <jh@suse.cz>
7641 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7642 flags; reorder rest of fields in more consistent way.
7643 (varpool_node): Remove analyzed, finalized and alias.
7644 (cgraph_ndoe): Likewise.
7645 (symtab_alias_ultimate_target): New function.
7646 (cgraph_function_node): Move offline.
7647 (cgraph_reset_node): Declare.
7648 (cgraph_comdat_can_be_unshared_p): Remove.
7649 (varpool_remove_initializer): Declare.
7650 (varpool_first_defined_variable, varpool_next_defined_variable
7651 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7652 (cgraph_function_with_gimple_body_p): Update.
7653 (varpool_all_refs_explicit_p): Update.
7654 (symtab_alias_target): New function.
7655 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7656 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7657 (cgraph_function_or_thunk_node): Simplify using
7658 symtab_alias_ultimate_target.
7659 (varpool_variable_node): Likewise.
7660 * cgraph.c (cgraph_create_function_alias): Update.
7661 (cgraph_add_thunk): Update.
7662 (cgraph_remove_node): Update.
7663 (dump_cgraph_node): Do not dump removed flags.
7664 (cgraph_function_body_availability): Update.
7665 (cgraph_propagate_frequency): Update.
7666 (verify_cgraph_node): Check sanity of local flag.
7667 (cgraph_function_node): Move here from cgraph.h; revamp for
7668 cgraph_function_or_thunk_node.
7669 * lto-symtab.c (lto_varpool_replace_node): Update.
7670 (lto_symtab_resolve_can_prevail_p): Update.
7671 (lto_symtab_merge_cgraph_nodes): Update.
7672 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7673 propagate_constants_accross_call, devirtualization_time_bonus,
7674 ipcp_propagate_stage): Update.
7675 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7676 * ipa-inline-transform.c (clone_inlined_nodes,
7677 preserve_function_body_p): Update.
7678 * ipa-reference.c (propagate): Update.
7679 (write_node_summary_p): Update.
7680 * toplev.c (wrapup_global_declaration_2): Update.
7681 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7682 (analyze_function) ... this one.
7683 (cgraph_process_new_functions): Update.
7684 (cgraph_reset_node): Export.
7685 (cgraph_finalize_function): Update.
7686 (cgraph_add_new_function): Update.
7687 (process_function_and_variable_attributes): Update.
7688 (varpool_finalize_decl): Update.
7689 (symbol_finalized): Remove.
7690 (symbol_finalized_and_needed): Rename to ...
7691 (symbol_defined_and_needed): ... update.
7692 (cgraph_analyze_functions): Update.
7693 (handle_alias_pairs): Update.
7694 (mark_functions_to_output): Update.
7695 (assemble_thunk): Update.
7696 (output_in_order): Update.
7697 (output_weakrefs): Update.
7698 (finalize_compilation_unit): Update.
7699 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7700 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7701 input_node, input_varpool_node): Update.
7702 * dbxout.c (dbxout_expand_expr): Update.
7703 * cgraphclones.c (cgraph_clone_node): Update.
7704 (cgraph_copy_node_for_versioning): Update.
7705 (cgraph_materialize_clone): Update.
7706 (cgraph_materialize_all_clones): Update.
7707 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7708 propagate_pure_const, propagate_nothrow): Update.
7709 * lto-streamer-out.c (lto_output, write_symbol): Update.
7710 * ipa-utils.c (ipa_reverse_postorder): Update.
7711 * ipa-inline.c (can_inline_edge_p): Update.
7712 (update_caller_keys, ipa_inline): Update.
7713 * dwarf2out.c (reference_to_unused,
7714 premark_types_used_by_global_vars_helper): Update.
7715 * tree-eh.c (tree_could_trap_p): Update.
7716 * ipa-split.c (consider_split, execute_split_functions): Update.
7717 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7718 has_addr_references_p): Update; move ahead in file for better
7720 (process_references): Simplify.
7721 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7723 (cgraph_comdat_can_be_unshared_p): Make static.
7724 (cgraph_externally_visible_p): Update.
7725 (varpool_externally_visible_p): Update.
7726 (function_and_variable_visibility): Update.
7727 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7728 ipa_tm_mark_force_output_node): Update.
7729 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7730 estimate_edge_devirt_benefit, inline_generate_summary,
7731 inline_write_summary): Update.
7732 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7733 * ipa-prop.c (ipa_compute_jump_functions): Update.
7734 (ipa_print_node_params, ipa_prop_read_section,
7735 ipa_update_after_lto_read, read_replacements_section): Update.
7736 * varasm.c (mark_decl_referenced): Update.
7737 (assemble_alias, dump_tm_clone_pairs): Update.
7738 * tree-inline.c (copy_bb): Update.
7739 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7741 * symtab.c (dump_symtab_base): Print new flags.
7742 (verify_symtab_base): Verify new flags.
7743 (symtab_alias_ultimate_target): New function.
7744 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7745 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7747 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7749 * i386.c (ix86_get_function_versions_dispatcher,
7750 ix86_generate_version_dispatcher_body): Update.
7751 (fold_builtin_cpu): Use varpool_add_new_variable.
7752 * varpool.c (varpool_remove_initializer): Break out from ...
7753 (varpool_remove_node): ... this one.
7754 (dump_varpool_node, varpool_node_for_asm,
7755 cgraph_variable_initializer_availability, varpool_analyze_node,
7756 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7757 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7760 2013-05-29 Jan Hubicka <jh@suse.cz>
7762 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7764 2013-05-29 Easwaran Raman <eraman@google.com>
7766 PR tree-optimization/57442
7767 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7768 when control exits the main loop.
7770 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7772 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7774 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7775 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7776 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7779 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7781 PR tree-optimization/57441
7782 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7783 Don't limit size of incr_vec to number of candidates.
7785 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7787 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7788 and mips16 directories.
7789 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7790 (MULTILIB_DIRNAMES): Ditto.
7791 (MULTILIB_EXCEPTIONS): Add new exceptions.
7792 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7793 (MULTILIB_DIRNAMES): Ditto.
7794 (MULTILIB_EXCEPTIONS): Add new exceptions.
7796 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7797 Marcus Shawcroft <marcus.shawcroft@arm.com>
7799 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7800 SYMBOL_TINY_ABSOLUTE.
7801 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7802 SYMBOL_TINY_ABSOLUTE.
7803 (aarch64_expand_mov_immediate): Likewise.
7804 (aarch64_classify_symbol): Likewise.
7805 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7806 Permit SYMBOL_TINY_ABSOLUTE.
7807 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7809 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7810 Marcus Shawcroft <marcus.shawcroft@arm.com>
7812 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7813 Refactor if/switch. Replace gcc_assert with if.
7815 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7817 * config/i386/i386.c (initial_ix86_tune_features): Enable
7818 FP Reassociation for AMD bdver1 and bdver2.
7820 2013-05-29 Martin Jambor <mjambor@suse.cz>
7822 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7823 and IMAGPART_EXPR do not occur within other handled_components.
7825 2013-05-29 Richard Biener <rguenther@suse.de>
7827 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7828 access on whether the use is in the BB we currently try to
7830 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7831 vectorize to vect_bb_slp_scalar_cost.
7833 2013-05-29 Richard Biener <rguenther@suse.de>
7835 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7836 computing scalar cost offsetted by stmts that are kept live
7838 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7839 for computation of scalar cost.
7841 2013-05-28 Steve Ellcey <sellcey@mips.com>
7843 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7845 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7847 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7848 array notation built-in reduction functions.
7849 * doc/passes.texi (Passes): Added documentation about changes done
7851 * doc/invoke.texi (C Dialect Options): Added documentation about
7852 the -fcilkplus flag.
7853 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7854 (BUILTINS_DEF): Depend on cilkplus.def.
7855 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7856 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7857 * cilkplus.def: New file.
7859 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7861 PR rtl-optimization/57439
7862 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7864 2013-05-28 Easwaran Raman <eraman@google.com>
7866 PR tree-optimization/57337
7867 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7868 (find_insert_point): Correctly identify the insertion point
7869 when two statements with the same UID is compared.
7871 2013-05-28 Richard Biener <rguenther@suse.de>
7873 PR tree-optimization/56787
7874 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7875 from the list of data references.
7876 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7878 (vect_analyze_loop_operations): Likewise.
7879 (vect_transform_loop): Remove clobbers.
7881 2013-05-28 Martin Jambor <mjambor@suse.cz>
7883 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7884 and REALPART_EXPRs have scalar type.
7886 2013-05-28 Richard Biener <rguenther@suse.de>
7888 PR tree-optimization/57411
7889 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7891 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7892 virtual operand propagation.
7894 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7896 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7897 destination register for bmasksi_vis.
7898 (vector_init_bshuffle): Likewise.
7899 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7901 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7903 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7904 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7905 mode if the instruction isn't available in the original mode.
7906 * config/sparc/sparc.opt (mfix-ut699): New option.
7907 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7908 (divdf3): Turn into expander.
7909 (divdf3_nofix): New insn.
7910 (divdf3_fix): Likewise.
7911 (divsf3): Disable if -mfix-ut699.
7912 (sqrtdf2): Turn into expander.
7913 (sqrtdf2_nofix): New insn.
7914 (sqrtdf2_fix): Likewise.
7915 (sqrtsf2): Disable if -mfix-ut699.
7917 2013-05-27 Richard Biener <rguenther@suse.de>
7920 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7921 block for the new loop.
7923 2013-05-27 Richard Biener <rguenther@suse.de>
7925 PR tree-optimization/57343
7926 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7927 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7928 (number_of_iterations_cond): Do not build the folded tree.
7930 2013-05-27 Richard Biener <rguenther@suse.de>
7934 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7935 OEP_CONSTANT_ADDRESS_OF retained.
7937 PR tree-optimization/57417
7938 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7940 (set_ssa_val_to): Compare addresses using
7941 get_addr_base_and_unit_offset.
7943 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7945 PR rtl-optimization/56833
7946 * postreload.c (move2add_record_mode): New function.
7947 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7948 (move2add_use_add2_insn): Use move2add_record_sym_value.
7949 (move2add_use_add3_insn): Likewise.
7950 (reload_cse_move2add): Use move2add_valid_value_p and
7951 move2add_record_mode. Invalidate call-clobbered and REG_INC
7952 affected regs by setting reg_mode to VOIDmode.
7953 (move2add_note_store): Don't pretend the inside of a SUBREG is
7954 the actual destination. Invalidate single/leading registers by
7955 setting reg_mode to VOIDmode.
7956 Use move2add_record_sym_value, move2add_valid_value_p and
7957 move2add_record_mode.
7959 2013-05-27 Richard Biener <rguenther@suse.de>
7961 PR tree-optimization/57396
7962 * tree-affine.c (double_int_constant_multiple_p): Properly
7963 return false for val == 0 and div != 0.
7965 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7967 * config/mips/mips.h: Use #elif in preprocessor conditions.
7969 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7972 * config/mips/constraints.md (kl): New constraint.
7973 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7974 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7975 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7977 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7979 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7982 * config/mips/mips.c (mips_can_inline_p): New function.
7983 (TARGET_CAN_INLINE_P): Define.
7985 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7987 * sched-int.h (ds_t, dw_t): Make unsigned int.
7988 Fix documentation that describes how all the ds_t bits are used.
7989 Reserve the last bit for delayed-branch scheduling.
7990 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7991 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7992 (gen_dep_weak_1): Remove prototype.
7993 * sched-deps.c (get_dep_weak_1): Make static.
7994 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7995 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7996 * doc/tm.texi: Regenerate.
7997 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7999 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8002 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8004 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8005 Sandra Loosemore <sandra@codesourcery.com>
8007 * config.gcc (powerpc-*): Allow native for with-cpu.
8009 2013-05-24 Jeff Law <law@redhat.com>
8011 PR tree-optimization/57124
8012 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8013 conversion feeding a condition if the range has an overflow
8014 if -fstrict-overflow. Add warnings for when we do make the
8017 2013-05-24 Dehao Chen <dehao@google.com>
8019 * tree-cfg.c (locus_discrim_map): Fix the typo.
8020 (locus_discrim_hasher): Likewise.
8021 (locus_discrim_hasher::hash): Likewise.
8022 (locus_discrim_hasher::equal): Likewise.
8024 2013-05-24 Martin Jambor <mjambor@suse.cz>
8026 PR tree-optimization/57294
8027 * cgraph.h (ipa_record_stmt_references): Declare.
8028 * cgraphbuild.c (ipa_record_stmt_references): New function.
8029 (build_cgraph_edges): Use ipa_record_stmt_references.
8030 (rebuild_cgraph_edges): Likewise.
8031 (cgraph_rebuild_references): Likewise.
8032 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8033 associated with the old statement and build references from the
8034 newly built statements.
8035 * ipa-ref.c (ipa_remove_stmt_references): New function.
8036 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8038 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8040 * lra-constraints.c (emit_spill_move): Use smaller mode for
8042 (check_and_process_move): Consider mem-reg moves for secondary
8044 (curr_insn_transform): Don't lose insns emitted before for
8045 secondary memory moves.
8046 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8047 reg set up in the current insn.
8049 2013-05-24 Dehao Chen <dehao@google.com>
8051 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8053 (locus_descrim_hasher::equal): Likewise.
8054 (build_gimple_cfg): New discriminator assignment algorithm.
8055 (make_edges): Likewise.
8056 (next_discriminator_for_locus): Likewise.
8057 (same_line_p): Likewise.
8058 (assign_discriminators): Likewise.
8059 (make_cond_expr_edges): Likewise.
8060 (make_gimple_switch_edges): Likewise.
8061 (make_goto_expr_edges): Likewise.
8062 (make_gimple_asm_edges): Likewise.
8064 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8066 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8067 X format specifier to only display bottom 16 bits.
8068 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8069 immediate to match for operand 2, since it will be masked.
8071 2013-05-24 Richard Biener <rguenther@suse.de>
8073 PR tree-optimization/57287
8074 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8075 all SSA names that occur in abnormal PHIs.
8077 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8080 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8081 that index is not negative.
8083 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8085 PR rtl-optimization/55177
8086 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8087 (simplify_byte_swapping_operation): New.
8088 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8089 (simplify_relational_operation_1): Deal with BSWAP.
8091 2013-05-23 Richard Henderson <rth@redhat.com>
8094 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8095 (ix86_reorg): Call it.
8097 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8100 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8101 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8102 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8104 2013-05-23 Christian Bruel <christian.bruel@st.com>
8107 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8109 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8110 Marcus Shawcroft <marcus.shawcroft@arm.com>
8112 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8113 * config/aarch64/constraints.md (Usa): Remove.
8114 * doc/md.texi (AArch64 Usa): Remove.
8116 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8117 Marcus Shawcroft <marcus.shawcroft@arm.com>
8119 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8120 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8121 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8122 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8124 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8126 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8127 instruction (AdvSIMD).
8128 * config/aarch64/aarch64-builtins.c
8129 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8130 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8132 2013-05-23 Martin Jambor <mjambor@suse.cz>
8135 * tree.h (contains_bitfld_component_ref_p): Declare.
8136 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8137 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8139 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8140 not access a bit-field. Assert all final offsets are byte-aligned.
8142 2013-05-23 Richard Biener <rguenther@suse.de>
8144 PR tree-optimization/57380
8145 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8146 least one invariant or re-used load.
8147 * passes.c (init_optimization_passes): Move pass_phiprop before
8150 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8152 * config/aarch64/aarch64-simd.md
8153 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8155 2013-05-23 Richard Biener <rguenther@suse.de>
8158 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8159 OEP_CONSTANT_ADDRESS_OF retained.
8161 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8164 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8165 don't lower unit. Handle unit not being always BITS_PER_WORD.
8167 2013-05-23 Richard Biener <rguenther@suse.de>
8169 PR rtl-optimization/57341
8170 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8171 instead of true_dependence.
8173 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8175 * bb-reorder.c (branch_threshold): Make const.
8176 (exec_threshold): Ditto.
8178 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8179 Pat Haugen <pthaugen@us.ibm.com>
8180 Peter Bergner <bergner@vnet.ibm.com>
8182 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8183 documentation for the power8 crypto builtins.
8185 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8187 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8188 macros for defining power8 builtin functions.
8189 (BU_P8V_AV_2): Likewise.
8190 (BU_P8V_AV_P): Likewise.
8191 (BU_P8V_VSX_1): Likewise.
8192 (BU_P8V_OVERLOAD_1): Likewise.
8193 (BU_P8V_OVERLOAD_2): Likewise.
8194 (BU_CRYPTO_1): Likewise.
8195 (BU_CRYPTO_2): Likewise.
8196 (BU_CRYPTO_3): Likewise.
8197 (BU_CRYPTO_OVERLOAD_1): Likewise.
8198 (BU_CRYPTO_OVERLOAD_2): Likewise.
8199 (XSCVSPDP): Fix typo, point to the correct instruction.
8200 (VCIPHER): Add power8 crypto builtins.
8201 (VCIPHERLAST): Likewise.
8202 (VNCIPHER): Likewise.
8203 (VNCIPHERLAST): Likewise.
8204 (VPMSUMB): Likewise.
8205 (VPMSUMH): Likewise.
8206 (VPMSUMW): Likewise.
8207 (VPERMXOR_V2DI): Likewise.
8208 (VPERMXOR_V4SI: Likewise.
8209 (VPERMXOR_V8HI: Likewise.
8210 (VPERMXOR_V16QI: Likewise.
8211 (VSHASIGMAW): Likewise.
8212 (VSHASIGMAD): Likewise.
8214 (VPERMXOR): Likewise.
8215 (VSHASIGMA): Likewise.
8217 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8218 __CRYPTO__ if the crypto instructions are available.
8219 (altivec_overloaded_builtins): Add support for overloaded power8
8222 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
8223 support for power8 crypto builtins.
8224 (builtin_function_type): Likewise.
8225 (altivec_init_builtins): Add support for builtins that take vector
8226 long long (V2DI) arguments.
8228 * config/rs6000/crypto.md: New file, define power8 crypto
8231 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8232 Pat Haugen <pthaugen@us.ibm.com>
8233 Peter Bergner <bergner@vnet.ibm.com>
8235 * doc/invoke.texi (Option Summary): Add power8 options.
8236 (RS/6000 and PowerPC Options): Likewise.
8238 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
8239 constraints.md instead of rs6000.h. Reorder w* constraints. Add
8240 wm, wn, wr documentation.
8242 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
8243 registers if direct move instructions are enabled.
8244 (wn): New constraint for no registers.
8245 (wq): New constraint for quad word even GPR registers.
8246 (wr): New constraint if 64-bit instructions are enabled.
8247 (wv): New constraint if power8 vector instructions are enabled.
8248 (wQ): New constraint for quad word memory locations.
8250 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
8251 constraint for 0..15 for crypto instructions.
8252 (gpc_reg_operand): If VSX allow registers in VSX registers as well
8253 as GPR and floating point registers.
8254 (int_reg_operand): New predicate to match only GPR registers.
8255 (base_reg_operand): New predicate to match base registers.
8256 (quad_int_reg_operand): New predicate to match even GPR registers
8257 for quad memory operations.
8258 (vsx_reg_or_cint_operand): New predicate to allow vector logical
8259 operations in both GPR and VSX registers.
8260 (quad_memory_operand): New predicate for quad memory operations.
8261 (reg_or_indexed_operand): New predicate for direct move support.
8263 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
8264 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
8265 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
8266 (POWERPC_MASKS): Add power8 options.
8267 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
8270 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8271 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
8273 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
8274 (-mpower8-fusion): New power8 options.
8275 (-mpower8-fusion-sign): Likewise.
8276 (-mpower8-vector): Likewise.
8277 (-mcrypto): Likewise.
8278 (-mdirect-move): Likewise.
8279 (-mquad-memory): Likewise.
8281 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
8283 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
8285 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
8286 (rs6000_debug_vector_unit): Add p8_vector.
8287 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
8288 definitions. Also print fusion state.
8289 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
8290 (rs6000_builtin_mask_calculate): Add power8 builtin support.
8291 (rs6000_option_override_internal): Add support for power8.
8292 (rs6000_common_init_builtins): Add debugging for skipped builtins
8294 (rs6000_adjust_cost): Add power8 support.
8295 (rs6000_issue_rate): Likewise.
8296 (insn_must_be_first_in_group): Likewise.
8297 (insn_must_be_last_in_group): Likewise.
8298 (force_new_group): Likewise.
8299 (rs6000_register_move_cost): Likewise.
8300 (rs6000_opt_masks): Likewise.
8302 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
8303 power8 capable assembler, default to power7 options.
8304 (TARGET_DIRECT_MOVE): Likewise.
8305 (TARGET_CRYPTO): Likewise.
8306 (TARGET_P8_VECTOR): Likewise.
8307 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
8308 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
8309 (VECTOR_MEM_P8_VECTOR_P): Likewise.
8310 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
8311 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
8312 (TARGET_XSCVDPSPN): Likewise.
8313 (TARGET_XSCVSPDPN): Likewsie.
8314 (TARGET_SYNC_HI_QI): Likewise.
8315 (TARGET_SYNC_TI): Likewise.
8316 (MASK_CRYPTO): Likewise.
8317 (MASK_DIRECT_MOVE): Likewise.
8318 (MASK_P8_FUSION): Likewise.
8319 (MASK_P8_VECTOR): Likewise.
8320 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
8321 temporary used by some of the direct move instructions to get two FP
8322 temporary registers does not force creation of a stack frame.
8323 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
8324 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
8325 that any VSX registers are tieable, even if they are also an
8326 Altivec vector mode.
8327 (r6000_reg_class_enum): Add wm, wr, wv constraints.
8328 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
8329 (RS6000_BTM_CRYPTO): Likewise.
8330 (RS6000_BTM_COMMON): Likewise.
8332 * config/rs6000/rs6000.md (cpu attribute): Add power8.
8333 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
8334 (enum rs6000_vector): Add power8 vector support.
8336 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8340 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
8341 (any_sibcall_could_use_r3): this and handle indirect calls.
8342 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
8344 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8346 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
8348 2013-05-22 Richard Biener <rguenther@suse.de>
8351 * profile.c (branch_prob): Do not split blocks that are
8352 abnormally receiving from ECF_RETURNS_TWICE functions.
8354 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8356 * recog.c (offsettable_address_addr_space_p): Fix calculation of
8357 address mode. Move pointer mode initialization to the same place.
8359 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8361 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
8362 while it has any effect.
8364 2013-05-21 Easwaran Raman <eraman@google.com>
8366 PR tree-optimization/57322
8367 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
8368 UID of the statement added to the BB to be 1.
8370 2013-05-21 Jakub Jelinek <jakub@redhat.com>
8372 PR tree-optimization/57331
8373 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
8374 of conversion from pointer type to integral type with integer.
8376 2013-05-21 Martin Jambor <mjambor@suse.cz>
8379 * ipa-prop.c (ipa_read_node_info): Process param_used and
8380 controlled_uses in the same order as when writing.
8382 2013-05-21 Magnus Granberg <baldrick@free.fr>
8385 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
8387 2013-05-21 Richard Biener <rguenther@suse.de>
8389 PR tree-optimization/57318
8390 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
8391 estimate stmts with side-effects as likely eliminated.
8393 2013-05-21 Richard Biener <rguenther@suse.de>
8395 PR tree-optimization/57330
8396 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
8397 preserve the call stmts fntype.
8399 2013-05-21 Richard Biener <rguenther@suse.de>
8401 PR tree-optimization/57303
8402 * tree-ssa-sink.c (statement_sink_location): Improve killing
8403 stmt detection and properly handle self-assignments.
8405 2013-05-21 Christian Bruel <christian.bruel@st.com>
8407 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
8408 spanning registers. LEAF_REG_REMAP is supported only for contiguous
8409 registers. Set register size out of the PARALLEL loop.
8411 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
8414 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
8415 (*fmasf4, *fmasf4_media): New insns.
8417 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8419 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
8420 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
8421 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
8422 (mips_idiv_insns): Update the comments to say that the returned
8423 instruction counts are in units of BASE_INSN_LENGTH.
8424 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
8425 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
8426 using 2 rather than 4 as the length of indirect MIPS16 and
8427 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
8428 length of a NOP. Don't divide MIPS16 lengths by 2.
8429 (mips16_split_long_branches): Assume a branch is long if the
8430 length is greater than 4 rather than 8.
8431 * config/mips/mips.md (length): Give MIPS16 lengths directly,
8432 rather than multiplying them by 2. Multiply instruction counts
8433 by BASE_INSN_LENGTH rather than 4.
8434 (*jump_mips16, tls_get_tp_mips16_<mode>)
8435 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
8437 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8439 * config/mips/mips.md (extended_mips16): Remove branch case.
8440 (length): Remove duplicated extended_mips16 test.
8442 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8444 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
8446 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
8448 * recog.h (Recog_data): Rename to...
8449 (recog_data_d): ...this.
8450 (recog_data): Update accordingly.
8451 * recog.c (recog_data): Likewise.
8452 * reload.c (save_recog_data): Likewise.
8453 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
8454 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
8456 2013-05-17 Julian Brown <julian@codesourcery.com>
8458 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
8459 found in a REG_EQUAL note, invalidate it.
8461 2013-05-17 Easwaran Raman <eraman@google.com>
8463 * tree-ssa-reassoc.c (find_insert_point): New function.
8464 (insert_stmt_after): Likewise.
8465 (get_def_stmt): Likewise.
8466 (ensure_ops_are_available): Likewise.
8467 (not_dominated_by): Likewise.
8468 (rewrite_expr_tree): Do not move statements beyond what is
8469 necessary. Remove call to swap_ops_for_binary_stmt...
8470 (reassociate_bb): ... and move it here.
8471 (build_and_add_sum): Assign UIDs for new statements.
8472 (linearize_expr): Likewise.
8473 (do_reassoc): Renumber gimple statement UIDs.
8475 2013-05-17 Jan Hubicka <jh@suse.cz>
8477 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8479 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8480 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8482 (output_weakrefs): Update.
8484 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8485 Martin Jambor <mjambor@suse.cz>
8488 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8489 value that corresponds to the given aggval is found in values vector.
8491 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8493 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8494 sse, sse2, sse3, ssse3 and sse4a flags to options.
8496 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8498 * gengtype-state.c: (s_expr_writer): New class, to handle
8499 prettifying of output layout of s-expressions.
8500 (state_writer): New class, to write out gtype.state.
8501 (state_written_type_count): Move this variable into member data of
8503 (s_expr_writer::s_expr_writer): New code: constructor for new class
8504 (state_writer::state_writer(): ditto
8505 (s_expr_writer::write_new_line): New function
8506 (s_expr_writer::write_any_indent): ditto
8507 (s_expr_writer::begin_s_expr): ditto
8508 (s_expr_writer::end_s_expr): ditto
8509 (write_state_fileloc): convert to method of state_writer...
8510 (state_writer:: write_state_fileloc): ...and use methods of
8511 s_expr_writer to write indentation into the gtype.state output file
8512 to visually represent the hierarchical structure of the list
8514 (write_state_fields): ditto, renaming to...
8515 (state_writer::write_state_fields)
8516 (write_state_a_string): ditto, renaming to...
8517 (state_writer::write_state_a_string)
8518 (write_state_string_option): ditto, renaming to...
8519 (state_writer::write_state_string_option)
8520 (write_state_type_option): ditto, renaming to...
8521 (state_writer::write_state_type_option)
8522 (write_state_nested_option): ditto, renaming to...
8523 (state_writer::write_state_nested_option)
8524 (write_state_option): ditto, renaming to...
8525 (state_writer::write_state_option)
8526 (write_state_options): ditto, renaming to...
8527 (state_writer::write_state_options)
8528 (write_state_lang_bitmap): ditto, renaming to...
8529 (state_writer::write_state_lang_bitmap)
8530 (write_state_version): ditto, renaming to...
8531 (state_writer::write_state_version)
8532 (write_state_scalar_type): ditto, renaming to...
8533 (state_writer::write_state_scalar_type)
8534 (write_state_string_type): ditto, renaming to...
8535 (state_writer::write_state_string_type)
8536 (write_state_undefined_type): ditto, renaming to...
8537 (state_writer::write_state_undefined_type)
8538 (write_state_struct_union_type): ditto, renaming to...
8539 (state_writer::write_state_struct_union_type)
8540 (write_state_struct_type): ditto, renaming to...
8541 (state_writer::write_state_struct_type)
8542 (write_state_user_struct_type): ditto, renaming to...
8543 (state_writer::write_state_user_struct_type)
8544 (write_state_lang_struct_type): ditto, renaming to...
8545 (state_writer::write_state_lang_struct_type)
8546 (write_state_param_struct_type): ditto, renaming to...
8547 (state_writer::write_state_param_struct_type)
8548 (write_state_pointer_type): ditto, renaming to...
8549 (state_writer::write_state_pointer_type)
8550 (write_state_array_type): ditto, renaming to...
8551 (state_writer::write_state_array_type)
8552 (write_state_gc_used): ditto, renaming to...
8553 (state_writer::write_state_gc_used)
8554 (write_state_common_type_content): ditto, renaming to...
8555 (state_writer::write_state_common_type_content)
8556 (write_state_type): ditto, renaming to...
8557 (state_writer::write_state_type)
8558 (write_state_pair_list): ditto, renaming to...
8559 (state_writer::write_state_pair_list)
8560 (write_state_pair): ditto, renaming to...
8561 (state_writer::write_state_pair)
8562 (write_state_typedefs): ditto, renaming to...
8563 (state_writer::write_state_typedefs)
8564 (write_state_structures): ditto, renaming to...
8565 (state_writer::write_state_structures)
8566 (write_state_param_structs): ditto, renaming to...
8567 (state_writer::write_state_param_structs)
8568 (write_state_variables): ditto, renaming to...
8569 (state_writer::write_state_variables)
8570 (write_state_srcdir): ditto, renaming to...
8571 (state_writer::write_state_srcdir)
8572 (write_state_files_list): ditto, renaming to...
8573 (state_writer::write_state_files_list)
8574 (write_state_languages): ditto, renaming to...
8575 (state_writer::write_state_languages)
8576 (write_state): create a state_writer instance and use it when
8577 writing out the state file
8579 2013-05-17 Mike Stump <mikestump@comcast.net>
8581 PR rtl-optimization/57304
8582 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8583 accessing DF_REF_REAL_LOC.
8585 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8587 PR rtl-optimization/57281
8588 PR rtl-optimization/57300
8589 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8590 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8591 what the other splitter did if the registers are dead.
8593 2013-05-17 Richard Biener <rguenther@suse.de>
8595 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8598 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8600 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8603 2013-05-17 Marek Polacek <polacek@redhat.com>
8605 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8606 length when doing non-zero store of storing '\0' to '\0'.
8608 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8610 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8611 vect_external_def oprnd1 with loop_vinfo, try to emit
8612 optional cast, negation and and stmts on the loop preheader
8613 edge instead of into the pattern def seq.
8615 PR tree-optimization/57051
8616 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8617 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8619 2013-05-16 Nick Clifton <nickc@redhat.com>
8621 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8622 (rl78_is_naked_func): New function.
8623 (rl78_expand_prologue): Skip prologue generation for naked functions.
8624 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8625 * doc/extend.texi (naked): Add RL78 to the list of processors
8626 that supports this attribute.
8628 2013-05-16 Jeff Law <law@redhat.com>
8630 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8632 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8634 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8635 cache parameters using detect_caches_amd also for CYRIX,
8636 NSC and TM2 signatures.
8638 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8639 Dzianis Kahanovich <mahatma@eu.by>
8643 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8644 VIA/Centaur processors and determine their cache parameters
8645 using detect_caches_amd.
8647 2013-05-16 Teresa Johnson <tejohnson@google.com>
8649 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8650 (rtl_verify_edges): New function.
8651 (rtl_verify_bb_insns): Ditto.
8652 (rtl_verify_bb_pointers): Ditto.
8653 (rtl_verify_bb_insn_chain): Ditto.
8654 (rtl_verify_fallthru): Ditto.
8655 (rtl_verify_bb_layout): Ditto.
8656 (rtl_verify_flow_info_1): Outline checks into new functions.
8657 (rtl_verify_flow_info): Ditto.
8659 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8661 * cfghooks.c (copy_bbs): Add update_dominance argument.
8662 * cfghooks.h (copy_bbs): Update prototype.
8663 * tree-cfg.c (gimple_duplicate_sese_region):
8664 Add update_dominance argument.
8665 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8666 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8667 gimple_duplicate_sese_region call.
8668 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8669 Update copy_bbs call.
8670 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8671 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8673 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8675 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8676 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8677 vect_recog_rotate_pattern.
8678 (vect_recog_rotate_pattern): New function.
8680 2013-05-16 Jason Merrill <jason@redhat.com>
8682 * Makefile.in (LLINKER): New variable.
8683 (mostlyclean): Remove link mutex.
8684 * configure.ac: Handle --enable-link-mutex.
8685 * lock-and-run.sh: New script.
8687 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8690 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8693 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8695 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8697 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8699 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8700 * config/arm/arm.c (next_consecutive_mem): New function.
8701 (gen_movmem_ldrd_strd): Likewise.
8702 * config/arm/arm.md (movmemqi): Update condition and code.
8703 (unaligned_loaddi, unaligned_storedi): New patterns.
8705 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8707 * config.gcc: Obsolete *-*-solaris2.9*.
8708 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8710 2013-05-16 Richard Biener <rguenther@suse.de>
8712 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8713 earlier, after GRAPHITE transforms and IV canonicalization.
8715 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8717 * omp-low.c (extract_omp_for_data): For collapsed loops,
8718 if at least one of the loops is known at compile time to
8719 iterate zero times, set count to 0.
8720 (expand_omp_regimplify_p): New function.
8721 (expand_omp_for_generic): For collapsed loops, if at least
8722 one of the loops isn't known to iterate at least once,
8723 add runtime check with setting count to 0.
8724 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8725 For unsigned types if it isn't known at compile time that
8726 the loop will iterate at least once, add runtime check to bypass
8727 the whole loop if initial condition isn't true.
8729 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8731 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8733 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8736 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8737 transformations to avoid an infinite loop.
8739 2013-05-16 Marek Polacek <polacek@redhat.com>
8741 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8743 2013-05-15 Leif Ekblad <leif@rdos.net>
8745 * config/i386/i386.c (ix86_decompose_address): Use
8746 DEFAULT_TLS_SEG_REG to access TLS segment register.
8747 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8748 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8749 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8751 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8754 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8755 sibling calls to functions that would normally be lazily bound,
8756 unless $gp is call-clobbered.
8758 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8760 * config/i386/i386.c (ix86_option_override_internal): Update
8761 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8762 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8763 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8764 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8766 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8768 2013-05-15 Andreas Schwab <schwab@suse.de>
8770 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8771 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8774 2013-05-15 Teresa Johnson <tejohnson@google.com>
8776 * loop-unroll.c (report_unroll_peel): Check decision before
8777 emitting unroll/peel message.
8779 2013-05-15 Teresa Johnson <tejohnson@google.com>
8781 * function.h (has_bb_partition): New rtl_data flag.
8782 (bb_reorder_complete): Ditto.
8783 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8784 instead of flag_reorder_blocks_and_partition.
8785 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8786 with some enhancements.
8787 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8788 * bb-reorder.c (connect_traces): Check for has_bb_partition
8789 instead of flag_reorder_blocks_and_partition.
8790 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8791 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8792 verify_hot_cold_block_grouping.
8793 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8795 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8798 * config/arm/predicates.md (call_insn_operand): New predicate.
8799 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8800 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8801 if insn is not a tail call.
8802 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8804 * config/arm/arm.h (enum reg_class): New caller save register class.
8805 (REG_CLASS_NAMES): Likewise.
8806 (REG_CLASS_CONTENTS): Likewise.
8807 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8810 2013-05-15 Richard Biener <rguenther@suse.de>
8812 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8813 of MSG_OPTIMIZED_LOCATIONS.
8814 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8815 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8817 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8818 of MSG_OPTIMIZED_LOCATIONS.
8819 (execute_vect_slp): Likewise.
8820 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8821 (vect_create_cond_for_alias_checks): Likewise.
8822 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8823 (vect_recog_widen_mult_pattern): Likewise.
8824 (vect_recog_widen_sum_pattern): Likewise.
8825 (vect_recog_over_widening_pattern): Likewise.
8826 (vect_recog_widen_shift_pattern): Likewise.
8827 (vect_recog_vector_vector_shift_pattern): Likewise.
8828 (vect_recog_divmod_pattern): Likewise.
8829 (vect_recog_mixed_size_cond_pattern): Likewise.
8830 (vect_recog_bool_pattern): Likewise.
8831 (vect_pattern_recog_1): Likewise.
8833 2013-05-15 Martin Jambor <mjambor@suse.cz>
8835 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8836 non-functions to builtin_unreachable.
8837 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8838 correct when new direct edges were discovered.
8840 2013-05-15 Martin Jambor <mjambor@suse.cz>
8842 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8843 header, print symbol order instead of node uid, print more information
8844 about indirect edge targets.
8845 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8847 (ipa_make_edge_direct_to_target): Likewise.
8848 (remove_described_reference): Likewise.
8849 (propagate_controlled_uses): Likewise.
8850 (ipa_print_node_params): Also print symbol order.
8851 (ipcp_transform_function): Print symbol order instead of node uids.
8852 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8853 (cgraph_get_create_real_symbol_node): Likewise.
8854 * ipa-cp.c (print_lattice): Likewise.
8855 (print_all_lattices): Likewise.
8856 (determine_versionability): Likewise.
8857 (initialize_node_lattices): Likewise.
8858 (estimate_local_effects): Likewise.
8859 (update_profiling_info): Likewise.
8860 (create_specialized_node): Likewise.
8861 (perhaps_add_new_callers): Likewise.
8862 (decide_about_value): Likewise.
8863 (decide_whether_version_node): Likewise.
8864 (identify_dead_nodes): Likewise.
8865 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8866 (dump_inline_summary): Likewise.
8867 (estimate_node_size_and_time): Likewise.
8868 (inline_analyze_function): Likewise.
8869 * ipa-inline.c (report_inline_failed_reason): Likewise.
8870 (want_early_inline_function_p): Likewise.
8871 (edge_badness): Likewise.
8872 (update_edge_key): Likewise.
8873 (inline_small_functions): Likewise. Add dumping of order to two other
8875 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8876 instead of node uids.
8877 (propagate_pure_const): Likewise.
8878 (propagate_pure_const): Likewise.
8879 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8880 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8881 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8883 * tree-pretty-print.c (dump_function_header): Likewise.
8884 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8885 Print symbol order instead of node uids.
8887 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8889 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8890 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8892 2013-05-15 Richard Biener <rguenther@suse.de>
8894 PR tree-optimization/57275
8895 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8896 return value for fail to do runtime alias checks for gather loads.
8898 2013-05-15 Jan Hubicka <jh@suse.cz>
8902 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8903 weakrefs are not external.
8904 (lto_symtab_merge_decls): Fix thinko when dealing with
8905 non-lto_symtab decls.
8906 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8907 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8908 * varpool.c (dump_varpool_node): Dump more flags.
8910 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8912 * config/i386/i386.c (processor_alias_table): Add instruction
8913 FSGSBASE for AMD bdver3 architecture.
8915 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8917 * tree.c (warn_deprecated_use): Print file:line using locus color.
8918 * diagnostic.c (diagnostic_report_current_module): Print file:line
8919 and file:line:column using locus color.
8921 2013-05-14 Mike Stump <mikestump@comcast.net>
8923 * gdbinit.in: Add __null.
8925 2013-05-14 Mike Stump <mikestump@comcast.net>
8927 * recog.h: Rename struct recog_data to Recog_data.
8928 * recog.c: Likewise.
8929 * reload.c (can_reload_into): Likewise.
8930 * config/picochip/picochip.c: Likewise.
8932 2013-05-14 Mike Stump <mikestump@comcast.net>
8934 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8936 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8938 * resource.h (struct resources): Remove unch_memory member.
8939 (CLEAR_RESOURCE): Don't clear unch_memory.
8940 * resource.c (mark_referenced_resources): Don't set it.
8941 (mark_set_resources): Likewise.
8942 (mark_target_live_regs): Don't clear it.
8943 (init_resource_info): Likewise.
8944 * reorg.c (resource_conflicts_p): Don't compare it.
8945 (redundant_insn): Don't set it.
8947 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8949 * emit-rtl.c (next_label): Remove unused function.
8950 (skip_consecutive_labels, link_cc0_insns): Move to ...
8951 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8952 only place where these functions are used, and make them static.
8954 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8956 * fold-const.c (fold_negate_expr): Handle vectors.
8957 (fold_truth_not_expr): Make it static.
8958 (fold_invert_truthvalue): New static function.
8959 (invert_truthvalue_loc): Handle vectors. Do not call
8960 fold_truth_not_expr directly.
8961 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8962 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8963 (fold_comparison): Handle vector constants.
8964 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8965 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8966 * tree.h (fold_truth_not_expr): Remove declaration.
8968 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8970 * config/aarch64/aarch64-simd.md
8971 (aarch64_vcond_internal<mode>): Rename to...
8972 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8973 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8974 float modes. Clarify all iterator modes.
8975 (vcond<mode><mode>): Use new name for vcond expanders.
8976 (vcond<v_cmp_result><mode>): Likewise.
8977 (vcondu<mode><mode>: Likewise.
8978 * config/aarch64/iterators.md (VDQF_COND): New.
8980 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8983 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8984 variable for the shift amount. Check that we shift by non-negative
8987 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8990 * config/arm/arm.h (EPILOGUE_USES): Only return true
8991 for LR_REGNUM after epilogue_completed.
8993 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8995 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8998 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9001 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9002 and Solaris 11+/x86 with gld.
9003 * configure: Regenerate.
9005 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9007 * expmed.c (expand_shift_1): Canonicalize rotates by
9008 constant bitsize / 2 to bitsize - 1.
9009 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9010 case ROTATERT>: Likewise.
9013 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9015 * config/i386/i386.md (rotateinv): New code attr.
9016 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9017 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9018 roll $31, %eax, etc.
9020 2013-05-14 Richard Biener <rguenther@suse.de>
9023 * tree-eh.c (sink_clobbers): Give up for successors with
9024 multiple predecessors and no virtual uses.
9026 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9028 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9029 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9031 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9034 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9035 the case when both op0 and op1 have VOIDmode.
9037 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9039 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9040 in multiply-accumulate mode.
9042 2013-05-13 Guozhi Wei <carrot@google.com>
9044 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9046 2013-05-13 Kai Tietz <ktietz@redhat.com>
9049 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9050 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9051 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9052 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9053 (ix86_expand_prologue): Likewise.
9054 (ix86_expand_split_stack_prologue): Likewise.
9055 (legitimate_pic_address_disp_p): Likewise.
9056 (legitimize_pic_address): Likewise.
9057 (legitimize_tls_address): Likewise.
9058 (legitimize_pe_coff_symbol): Likewise.
9059 (output_pic_addr_const): Likewise.
9060 (construct_plt_address): Likewise.
9061 (ix86_expand_call): Likewise.
9062 (x86_output_mi_thunk): Likewise.
9063 (x86_function_profiler): Likewise.
9065 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9067 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9068 similar switch cases.
9069 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9070 (aarch64_simd_mov_to_<mode>low): Delete.
9071 (aarch64_simd_mov_to_<mode>high): Delete.
9072 (move_lo_quad_<mode>): Add w<-r alternative.
9073 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9074 (aarch64_simd_mov_from_*): Update type attribute.
9075 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9078 2013-05-13 Jan Hubicka <jh@suse.cz>
9080 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9081 * config/i386/i386.c (ix86_compute_frame_layout,
9082 ix86_expand_epilogue, emit_i387_cw_initialization,
9083 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9084 ix86_local_alignment): Fix use of size/speed predicates.
9086 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9088 PR tree-optimization/45216
9089 PR tree-optimization/57157
9090 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9091 the (-Y) & (B - 1) variant if OP is |.
9092 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9093 return shifted. Use (-op1) & (prec - 1) as other_amount
9094 instead of prec - op1.
9096 2013-05-13 Martin Jambor <mjambor@suse.cz>
9099 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9100 (ipa_constant_data): New type.
9101 (ipa_jump_func): Use ipa_constant_data to hold information about
9102 constant jump functions.
9103 (ipa_get_jf_constant): Adjust to jump function type changes.
9104 (ipa_get_jf_constant_rdesc): New function.
9105 (ipa_param_descriptor): New field controlled_uses.
9106 (ipa_get_controlled_uses): New function.
9107 (ipa_set_controlled_uses): Likewise.
9108 * ipa-ref.h (ipa_find_reference): Declare.
9109 * ipa-prop.c (ipa_cst_ref_desc): New type.
9110 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9112 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9113 New parameter cs. Adjust all callers.
9114 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9115 (remove_described_reference): New function.
9116 (jfunc_rdesc_usable): Likewise.
9117 (try_make_edge_direct_simple_call): Decrement controlled use count,
9118 attempt to remove reference if it hits zero.
9119 (combine_controlled_uses_counters): New function.
9120 (propagate_controlled_uses): Likewise.
9121 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9122 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9123 (ipa_print_node_params): Print described use counter.
9124 (ipa_write_jump_function): Adjust to jump function type changes.
9125 (ipa_read_jump_function): New parameter CS, pass it to
9126 ipa_set_jf_constant. Adjust caller.
9127 (ipa_write_node_info): Stream controlled use count
9128 (ipa_read_node_info): Likewise.
9129 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9131 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9132 count. Remove cloning-added reference if it reaches zero.
9133 * ipa-ref.c (ipa_find_reference): New function.
9135 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9137 * config/i386/i386.c (processor_target_table): Modified default
9138 alignment values for AMD BD and BT architectures.
9140 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9142 * tree-vect-generic.c (uniform_vector_p): Move ...
9143 * tree.c (uniform_vector_p): ... here.
9144 * tree.h (uniform_vector_p): Declare it.
9145 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9148 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9150 PR tree-optimization/57230
9151 * tree-ssa-strlen.c (handle_char_store): Record length for
9152 array store from STRING_CST.
9154 PR tree-optimization/57230
9155 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9158 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9160 * config/epiphany/epiphany.c (epiphany_init): Check size of
9161 NUM_MODES_FOR_MODE_SWITCHING.
9162 (epiphany_expand_prologue):
9163 Remove CONFIG_REGNUM initial value handling code.
9164 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9165 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9166 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9167 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9168 Don't return 1 for FP_MODE_NONE.
9169 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9170 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9171 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9172 * config/epiphany/epiphany.md (save_config): New pattern.
9174 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9176 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9178 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9180 * config/i386/i386.md (memory): Handle sseishft1.
9181 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9182 (*vec_extractv2di_1): Ditto.
9184 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9186 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9189 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9191 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9192 Add mthumb/march=armv7-a multilib.
9193 Add mthumb/march=armv7-r multilib.
9194 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
9196 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
9198 * config/v850/t-rtems: Add more multilibs.
9200 2013-05-10 Richard Biener <rguenther@suse.de>
9202 PR tree-optimization/57214
9203 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
9204 not propagate from SSA names that occur in abnormal PHI nodes.
9206 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
9208 * stor-layout.c (element_precision): New function.
9209 * machmode.h (element_precision): Declare it.
9210 * tree.c (build_minus_one_cst): New function.
9211 (element_precision): Likewise.
9212 * tree.h (build_minus_one_cst): Declare new function.
9213 (element_precision): Likewise.
9214 * fold-const.c (operand_equal_p): Use element_precision.
9215 (fold_binary_loc): Handle vector types.
9216 * convert.c (convert_to_integer): Use element_precision.
9217 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
9220 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
9222 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
9223 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
9224 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
9225 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9226 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
9227 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
9228 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
9229 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9230 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
9231 (Uuw8): New constraints.
9232 (Usb4): Move into alphabetical order.
9233 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
9234 (sd8_operand, ub8_operand, uw8_operand): New predicates.
9235 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
9236 previously unnamed patterns.
9237 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
9238 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
9239 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
9240 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
9241 of set_attr_alternative/if_then_else. Use extended_mips16 instead
9242 of specific lengths.
9244 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9246 * config/i386/i386.md (rotateinv): New code attr.
9247 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9248 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9249 roll $31, %eax, etc.
9251 PR tree-optimization/45216
9252 PR tree-optimization/57157
9253 * tree-ssa-forwprop.c (simplify_rotate): New function.
9254 (ssa_forward_propagate_and_combine): Call it.
9256 2013-05-10 Richard Biener <rguenther@suse.de>
9258 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
9259 disable peeling when we version for aliasing.
9260 (vector_alignment_reachable_p): Honor explicit user alignment.
9261 (vect_supportable_dr_alignment): Likewise.
9262 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
9263 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
9264 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
9265 then peeling to arrange for the cost-model check to come first.
9267 2013-05-10 Alan Modra <amodra@gmail.com>
9269 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
9270 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
9271 * configure: Regenerate.
9273 2013-05-10 Alan Modra <amodra@gmail.com>
9276 * varasm.c (default_elf_select_section): Move !DECL_P check..
9277 (get_named_section): ..to here before calling get_section_name.
9279 (default_section_type_flags): Add DECL_P check.
9280 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
9281 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
9283 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
9285 * config/epiphany/epiphany.c (epiphany_expand_prologue):
9286 When using gen_stack_adjust_str with a register offset, add a
9287 REG_FRAME_RELATED_EXPR note.
9289 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
9291 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
9292 (*vec_extractv4si_zext_mem): Ditto.
9293 (*vec_extractv2di): Add 0->x and x->x alternatives.
9294 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
9295 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
9297 2013-05-09 Jason Merrill <jason@redhat.com>
9299 N3639 C++1y VLA support
9300 * gimplify.c (gimplify_vla_decl): Don't touch an existing
9303 * tree.c (build_constructor_va): New.
9304 * tree.h: Declare it.
9306 2013-05-09 Martin Jambor <mjambor@suse.cz>
9309 * gimple-fold.c (canonicalize_constructor_val): Call
9310 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
9312 2013-05-09 Jan Hubicka <jh@suse.cz>
9313 Richard Biener <rguenther@suse.de>
9316 * symtab.c (symtab_make_decl_local): Do not add private names.
9318 2013-05-09 Jan Hubicka <jh@suse.cz>
9321 * symtab.c (insert_to_assembler_name_hash): Handle clones.
9322 (unlink_from_assembler_name_hash): Likewise.
9323 (symtab_prevail_in_asm_name_hash, symtab_register_node,
9324 symtab_unregister_node, symtab_initialize_asm_name_hash,
9325 change_decl_assembler_name): Update.
9327 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
9329 * config/aarch64/aarch64.md: New movtf split.
9330 (*movtf_aarch64): Update.
9331 (aarch64_movdi_tilow): Handle TF modes and rename to
9332 aarch64_movdi_<mode>low.
9333 (aarch64_movdi_tihigh): Handle TF modes and rename to
9334 aarch64_movdi_<mode>high
9335 (aarch64_movtihigh_di): Handle TF modes and rename to
9336 aarch64_mov<mode>high_di
9337 (aarch64_movtilow_di): Handle TF modes and rename to
9338 aarch64_mov<mode>low_di
9339 (aarch64_movtilow_tilow): Remove spurious whitespace.
9340 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
9342 (aarch64_print_operand): Update.
9344 2013-05-09 Alan Modra <amodra@gmail.com>
9346 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
9348 * configure: Regenerate.
9350 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9352 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
9353 splitter preparation statements.
9354 * config/i386/sse.md (*vec_extract* splitters): Ditto.
9355 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
9358 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9360 * gimple-ssa-strength-reduction.c (count_candidates): Change
9361 return value to int.
9362 (analyze_candidates_and_replace): Change type of length to int.
9364 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9366 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
9367 (*vec_extract<mode>): Use VI12_128 mode iterator.
9368 (*vec_extract<mode>_mem): Ditto.
9369 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
9372 2013-05-08 Diego Novillo <dnovillo@google.com>
9377 2012-08-17 Diego Novillo <dnovillo@google.com>
9380 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
9381 * config.in: Regenerate.
9382 * configure: Regenerate.
9383 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
9385 2013-05-08 Jan Hubicka <jh@suse.cz>
9388 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
9389 * cgraph.h (symtab_node_base): Add unique_name.
9390 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9391 input_overwrite_node, input_varpool_node): Stream unique_name.
9392 * cgraphclones.c (cgraph_create_virtual_clone,
9393 cgraph_function_versioning): Set unique_name.
9394 * ipa.c (function_and_variable_visibility): Set unique_name.
9396 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9398 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
9399 (alloc_cand_and_find_basis): Restrict conditional candidate
9400 processing to CAND_MULTs.
9402 2013-05-08 Jan Hubicka <jh@suse.cz>
9405 lto-symtab.c (lto_symtab_symbol_p): New function.
9406 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
9407 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
9408 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
9409 Skip static symbols.
9411 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
9413 PR tree-optimization/57200
9414 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
9415 Only call inform if the preceding warning_at returns true.
9417 2013-05-07 Han Shen <shenhan@google.com>
9419 * cfgexpand.c (record_or_union_type_has_array_p): New function.
9420 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
9421 * common.opt (fstack-protector-strong): New option.
9422 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
9423 * doc/invoke.texi (Optimization Options): Document
9424 "-fstack-protector-strong".
9425 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
9427 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9429 * config/mips/mips.c (mips_machine_reorg2): Return 0.
9431 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
9433 * ira.c (update_equiv_regs): Add insn having equiv memory even if
9434 it is not lhs of the insn.
9435 (setup_reg_equiv): Remove insn having equiv memory which it is not
9437 * lra-constraints.c (process_address): Try to improve generation
9438 code for address base + disp.
9439 (lra_constraints): Make correct the code for checking insn setting
9440 up backward equivalence. Remove insn only if it is in the init
9442 * lra-eliminations.c (update_reg_eliminate): Change return value.
9443 (lra_eliminate): Use the result.
9445 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
9447 * config/i386/sse.md (ssescalarnummask): New mode attribute.
9448 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
9449 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
9450 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
9451 register target operands.
9452 (*vec_extractv8hi_sse2): New pattern.
9453 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
9454 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
9455 (*vec_extract<mode>_mem): New insn and split pattern.
9457 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
9459 * config/arm/arm.c (arm_asan_shadow_offset): New function.
9460 (TARGET_ASAN_SHADOW_OFFSET): Define.
9461 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9462 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9464 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9466 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9467 (incr_vec_index): Return -1 if increment not found.
9468 (create_add_on_incoming_edge): Assert if increment not found.
9469 (record_increment): Limit number of increments recorded.
9470 (all_phi_incrs_profitable): Return false if an increment not found.
9471 (replace_profitable_candidates): Don't process increments that were
9473 (analyze_candidates_and_replace): Limit size of incr_vec.
9475 2013-05-07 Richard Biener <rguenther@suse.de>
9477 * calls.c (special_function_p): setjmp-like functions are leaf.
9478 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9479 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9481 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9483 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9484 (aarch64_simd_mov<mode>): New expander.
9485 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9486 (aarch64_simd_mov_to_<mode>high): Likewise.
9487 (aarch64_simd_mov_from_<mode>low): Likewise.
9488 (aarch64_simd_mov_from_<mode>high): Likewise.
9489 (aarch64_dup_lane<mode>): Update.
9490 (aarch64_dup_lanedi): New instruction pattern.
9491 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9492 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9494 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9496 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9497 (replace_mult_candidate): Remove unnecessary argument; remove
9498 unnecessary parameter from call to introduce_cast_before_cand.
9499 (replace_unconditional_candidate): Remove unnecessary parameter
9500 from call to replace_mult_candidate.
9501 (replace_conditional_candidate): Likewise.
9502 (insert_initializers): Use make_temp_ssa_name.
9503 (introduce_cast_before_cand): Remove unnecessary argument; use
9505 (replace_one_candidate): Remove unnecessary argument; remove
9506 unnecessary parameter from calls to introduce_cast_before_cand.
9507 (replace_profitable_candidates): Remove unnecessary parameters
9508 from calls to replace_one_candidate.
9510 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9512 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9513 phi def as possibly hiding a basis for a CAND_ADD whose operands
9514 have been commuted in the analysis.
9515 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9517 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9519 * config/aarch64/aarch64.md
9520 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9521 shift value between 0-4.
9523 2013-05-07 Richard Biener <rguenther@suse.de>
9525 * double-int.h (rshift): New overload.
9526 * double-int.c (rshift): New function.
9527 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9528 (create_reference_ops_from_ref): Remove.
9529 (vn_reference_insert): Use shared ops for constructing the
9530 reference and copy it.
9532 2013-05-07 Richard Biener <rguenther@suse.de>
9535 * tree-eh.c (sink_clobbers): Properly propagate
9536 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9538 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9540 PR tree-optimization/57149
9541 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9542 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9543 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9544 uninit_undefined_value_p instead of ssa_undefined_value_p.
9547 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9548 for modifier == EXPAND_INITIALIZER.
9550 2013-05-07 Anton Blanchard <anton@samba.org>
9552 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9553 for powerpc64 little endian.
9554 * configure: Regenerate.
9556 2013-05-06 Graham Stott <grahams@btinternet.com>
9558 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9559 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9560 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9561 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9563 2013-05-06 Graham Stott <grahams@btinternet.com>
9565 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9566 codes which allow non-lvalues.
9568 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9570 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9571 components are all 1s.
9572 (integer_minus_onep): New function.
9573 * tree.h (integer_minus_onep): Declare it.
9574 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9575 integer_minus_onep instead of integer_all_onesp.
9577 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9580 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9581 variations of these patterns.
9583 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9585 * config/i386/i386.md (isa): Add x64_sse4 member.
9586 (enabled): Handle x64_sse4.
9587 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9588 instruction for 64bit SSE4_1 targets. Update insn attributes.
9589 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9590 instruction for SSE4_1 targets. Update insn attributes.
9591 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9592 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9593 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9595 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9596 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9597 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9599 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9602 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9604 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9606 * final.c (do_assembler_dialects): Don't handle curly braces and
9607 vertical bar escaped by % as dialect delimiters.
9608 (output_asm_insn): Print curly braces and vertical bar if escaped
9609 by % and ASSEMBLER_DIALECT defined.
9610 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9611 * doc/tm.texi: Regenerated.
9613 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9615 * config/mips/mips.c: Include tree-pass.h.
9616 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9617 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9618 (pass_mips_machine_reorg2): New machine specific pass.
9619 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9620 (mips_option_override): Register the new pass.
9621 * rtl.h (cleanup_barriers): Remove prototype.
9622 (dbr_schedule): Likewise.
9623 * jump.c (cleanup_barriers): Make static.
9624 * reorg.c (dbr_schedule): Likewise.
9626 2013-05-06 Richard Biener <rguenther@suse.de>
9628 PR tree-optimization/57185
9629 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9632 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9635 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9636 between "first_arg" and "insn", not "last" and "insn".
9638 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9640 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9641 (find_candidates_in_block): Re-enable slsr_process_phi.
9642 (create_phi_basis): Fix double counting of candidate adjustment.
9644 2013-05-06 Richard Biener <rguenther@suse.de>
9647 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9648 the edge is also fallthru, preserve it and just clear the
9650 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9651 also complex, preserve that and just clear the fallthru flag.
9652 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9653 update virtual operands.
9655 2013-05-06 Alan Modra <amodra@gmail.com>
9657 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9658 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9659 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9660 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9661 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9662 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9664 2013-05-06 Alan Modra <amodra@gmail.com>
9666 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9667 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9668 (DEFAULT_ASM_ENDIAN): Define.
9669 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9670 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9671 Update -K PIC clause from sysv4.h.
9672 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9673 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9675 2013-05-06 Alan Modra <amodra@gmail.com>
9677 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9678 twice for little-endian.
9679 (ashrdi3_no_power, ashrdi3): Support little-endian.
9681 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9684 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9685 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9687 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9688 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9689 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9692 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9693 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9695 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9696 * config/pa/pa.opt: Make mbig-switch a no-op.
9697 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9698 (CASE_VECTOR_MODE): Always return SImode.
9699 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9700 for the !TARGET_BIG_SWITCH case.
9701 * config/pa/pa-linux.h: Likewise.
9702 * config/pa/pa-openbsd.h: Likewise.
9703 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9704 * config/pa/pa.md (short_jump): Remove define_insn.
9705 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9706 (casesi0): Remove define_insn.
9707 (type): Remove btable_branch.
9708 (pa_combine_type): Likewise.
9709 (in_nullified_branch_delay): Likewise.
9710 (in_call_delay): Likewise.
9711 (define_delay): Likewise.
9712 (define_insn_reservation "Z3"): Likewise.
9713 (define_insn_reservation "Z4"): Likewise.
9714 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9715 (pa_adjust_insn_length): Remove adjustment for btable branches.
9716 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9719 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9721 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9722 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9723 Add m->r,x alternatives.
9724 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9725 splitters using SWI48x mode iterator.
9726 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9727 TARGET_64BIT. Add m->x alternative.
9728 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9729 Add o->x alternative. Enable for TARGET_SSE.
9730 (sse_storeq): Remove expander.
9731 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9732 with memory input operand.
9733 (*vec_extractv2di_1 splitter): New.
9734 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9735 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9737 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9739 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9740 (INT_LOWPART): Delete.
9741 (extract_MB): Adjust.
9742 (extract_ME): Adjust.
9743 (print_operand): Adjust.
9745 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9747 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9748 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9749 (reg_or_logical_cint_operand, easy_fp_constant,
9750 logical_const_operand): Delete "CONST_DOUBLE" case.
9751 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9752 "HOST_BITS_PER_WIDE_INT == 64" test.
9753 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9754 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9755 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9756 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9758 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9759 CONST_DOUBLE DImode/VOIDmode case.
9760 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9761 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9762 CONST_DOUBLE VOIDmode case.
9763 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9764 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9765 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9766 Delete CONST_DOUBLE case.
9767 (splitters for mov FMOVE64 const_double): Delete
9768 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9769 "HOST_BITS_PER_WIDE_INT >= 64" test.
9770 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9772 (mov DI const_double): Delete.
9774 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9776 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9777 on op shows all bits zero in mode of a lowpart subreg, return zero.
9779 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9782 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9783 to save TFmode registers and DImode to save TImode registers for
9784 caller save operations.
9785 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9786 mark being partially clobbered since they only use the first
9789 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9790 and TDmode only use the upper 64-bits of each VSX register.
9792 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9794 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9795 (find_candidates_in_block): Disable slsr_process_phi.
9797 2013-05-03 Guozhi Wei <carrot@google.com>
9799 * coverage.c (coverage_obj_init): Move the construction of gcov
9801 (build_init_ctor): ... here.
9803 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9805 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9806 (slsr_cand_d): Redefine def_phi.
9807 (stride_status, phi_adjust_status, count_phis_status): New enums.
9808 (find_phi_def): New.
9809 (find_basis_for_base_expr): New.
9810 (find_basis_for_candidate): Handle hidden bases.
9811 (alloc_cand_and_find_basis): Handle phi candidates.
9812 (slsr_process_phi): New.
9813 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9814 (create_mul_imm_cand): Likewise.
9815 (create_add_ssa_cand): Exclude phi base candidates.
9816 (create_add_imm_cand): Likewise.
9817 (slsr_process_cast): Likewise.
9818 (slsr_process_copy): Likewise.
9819 (find_candidates_in_block): Handle phi candidates.
9820 (dump_candidate): Likewise.
9821 (unconditional_cands): Delete.
9822 (unconditional_cands_with_known_stride_p): Delete.
9823 (phi_dependent_cand_p): New.
9824 (cand_increment): Handle phi-dependent candidates.
9825 (replace_dependent): Delete.
9826 (replace_mult_candidate): New.
9827 (replace_unconditional_candidate): New.
9828 (incr_vec_index): Move to avoid forward reference.
9829 (create_add_on_incoming_edge): New.
9830 (create_phi_basis): New.
9831 (replace_dependents): Delete.
9832 (replace_conditional_candidate): New.
9833 (phi_add_costs): New.
9834 (replace_uncond_cands_and_profitable_phis): New.
9835 (record_increment): Handle phi adjustments.
9836 (record_phi_increments): New.
9837 (record_increments): Handle phi adjustments.
9838 (phi_incr_cost): New.
9839 (lowest_cost_path): Handle phis.
9840 (total_savings): Likewise.
9841 (analyze_increments): Likewise.
9842 (ncd_with_phi): New.
9843 (ncd_of_cand_and_phis): New.
9844 (nearest_common_dominator_for_cands): Handle phi increments.
9845 (all_phi_incrs_profitable): New.
9846 (replace_profitable_candidates): Handle phi-dependent candidates.
9847 (analyze_candidates_and_replace): Likewise.
9849 2013-05-03 Teresa Johnson <tejohnson@google.com>
9852 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9853 do not exceed REG_BR_PROB_BASE.
9855 2013-05-03 Jeff Law <law@redhat.com>
9857 PR tree-optimization/57144
9858 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9859 operand of the condition will bit into the new type when eliminating
9860 a cast feeding a condition.
9862 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9864 PR rtl-optimization/57130
9865 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9866 of COMPARE as in_code to the recursive call if needed.
9868 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9870 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9871 (enabled): Handle new members.
9872 * config/i386/sse.md (*vec_concatv2si): Merge from
9873 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9874 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9876 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9878 PR tree-optimization/57027
9879 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9880 for fnms opportunity, check we got the prerequisite kind
9881 of tree / gimple before using accessor functions.
9883 2013-05-03 Richard Biener <rguenther@suse.de>
9885 * double-int.h (lshift): New overload without precision
9887 (operator *=, operator +=, operator -=): Move ...
9888 * double-int.c (operator *=, operator +=, operator -=): ... here
9889 and implement more efficiently.
9890 (mul_double_with_sign): Remove.
9891 (lshift_double): Adjust to take unsinged shift argument, push
9892 dispatching code to callers.
9893 (mul_double_wide_with_sign): Add early out for callers that
9894 are not interested in high parts or overflow.
9895 (lshift): New function.
9896 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9898 (lrotate, rrotate): Use logical shifts.
9899 * expr.c (get_inner_reference): Use lshift.
9900 * fixed-value.c (do_fixed_divide): Likewise.
9901 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9902 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9903 (indirect_refs_may_alias_p): Likewise.
9904 (stmt_kills_ref_p_1): Likewise.
9906 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9908 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9910 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9912 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9913 scalar form of FABD instruction.
9915 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9917 * lra-constraints.c (process_alt_operands): Add checking alt
9918 number to choose the best alternative.
9920 2013-05-02 Richard Biener <rguenther@suse.de>
9922 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9923 bitmap and its handling.
9924 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9926 2013-05-02 Richard Biener <rguenther@suse.de>
9929 * tree-inline.c (copy_loops): Properly handle removed loops.
9930 (copy_cfg_body): Mark destination loops for fixup if source
9933 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9936 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9937 generating simple_return for naked functions.
9939 2013-05-02 Martin Jambor <mjambor@suse.cz>
9942 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9943 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9945 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9946 ipa_agg_replacement_value structures.
9947 (known_aggs_to_agg_replacement_list): Likewise.
9948 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9949 (read_agg_replacement_chain): Likewise.
9950 (ipcp_transform_function): Also check that by_ref flags match.
9952 2013-05-02 Richard Biener <rguenther@suse.de>
9954 * graphds.h (struct graph): Add obstack member.
9955 * graphds.c (new_graph): Initialize obstack and allocate
9957 (add_edge): Allocate edge from the obstack.
9958 (free_graph): Free the obstack instead of all edges and vertices.
9960 2013-05-02 Teresa Johnson <tejohnson@google.com>
9962 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9964 * cfg.c (update_bb_profile_for_threading): Ditto.
9965 * tree-inline.c (copy_bb): Ditto.
9966 (copy_edges_for_bb): Ditto.
9967 (initialize_cfun): Ditto.
9968 (copy_cfg_body): Ditto.
9969 (expand_call_inline): Ditto.
9970 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9971 (estimate_node_size_and_time): Ditto.
9972 (inline_merge_summary): Ditto.
9973 * cgraphclones.c (cgraph_clone_edge): Ditto.
9974 (cgraph_clone_node): Ditto.
9975 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9976 (compute_trg_info): Ditto.
9978 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9980 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9981 S reg when fp attribute set.
9982 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9984 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9986 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9988 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9989 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9990 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9992 2013-05-02 Richard Biener <rguenther@suse.de>
9994 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9995 (struct instantiate_cache_entry): New type.
9996 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9997 (struct instantiate_cache_type): New type.
9998 (set_instantiated_value, get_instantiated_value): Remove.
9999 (get_instantiated_value_entry): New function.
10000 (instantiate_scev_name): Use the new cache and adjust.
10001 (instantiate_scev_poly): Adjust.
10002 (instantiate_scev_binary): Likewise.
10003 (instantiate_array_ref): Likewise.
10004 (instantiate_scev_convert): Likewise.
10005 (instantiate_scev_not): Likewise.
10006 (instantiate_scev_3): Likewise.
10007 (instantiate_scev_2): Likewise.
10008 (instantiate_scev_r): Likewise.
10009 (instantiate_scev): Likewise.
10010 (resolve_mixers): Likewise.
10012 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10015 * lra-constraints.c (best_small_class_operands_num): Remove.
10016 (process_alt_operands): Remove small_class_operands_num. Take
10017 small classes operands into losers and only if the operand is not
10018 matched. Modify debugging output.
10019 (curr_insn_transform): Remove best_small_class_operands_num.
10022 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10024 * config/aarch64/aarch64-builtins.c
10025 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10026 * config/aarch64/aarch64-simd-builtins.def
10027 (reduc_splus_): Add new modes.
10028 (reduc_uplus_): New.
10029 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10030 (reduc_uplus_v4sf): Likewise.
10031 (reduc_splus_v4sf): Likewise.
10032 (aarch64_addv<mode>): Likewise.
10033 (reduc_uplus_<mode>): Likewise.
10034 (reduc_splus_<mode>): Likewise.
10035 (aarch64_addvv2di): Likewise.
10036 (reduc_uplus_v2di): Likewise.
10037 (reduc_splus_v2di): Likewise.
10038 (aarch64_addvv2si): Likewise.
10039 (reduc_uplus_v2si): Likewise.
10040 (reduc_splus_v2si): Likewise.
10041 (reduc_<sur>plus_<mode>): New.
10042 (reduc_<sur>plus_v2di): Likewise.
10043 (reduc_<sur>plus_v2si): Likewise.
10044 (reduc_<sur>plus_v4sf): Likewise.
10045 (aarch64_addpv4sf): Likewise.
10046 * config/aarch64/arm_neon.h
10047 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10048 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10049 add UNSPEC_SADDV, UNSPEC_UADDV.
10051 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10053 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10055 * config/aarch64/arm_neon.h
10056 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10058 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10060 * config/aarch64/aarch64-builtins
10061 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10063 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10065 * config/aarch64/aarch64-simd-builtins.def
10066 (reduc_smax_): New.
10067 (reduc_smin_): Likewise.
10068 (reduc_umax_): Likewise.
10069 (reduc_umin_): Likewise.
10070 (reduc_smax_nan_): Likewise.
10071 (reduc_smin_nan_): Likewise.
10074 (smax): Update for V2SF, V4SF and V2DF modes.
10077 (smin_nan): Likewise.
10078 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10079 (<su><maxmin><mode>3): ...This, refactor.
10080 (s<maxmin><mode>3): New.
10081 (<maxmin_uns><mode>3): Likewise.
10082 (reduc_<maxmin_uns>_<mode>): Refactor.
10083 (reduc_<maxmin_uns>_v4sf): Likewise.
10084 (reduc_<maxmin_uns>_v2si): Likewise.
10085 (aarch64_<fmaxmin><mode>: Remove.
10086 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10088 (vmin<q>_f<32,64>): Likewise.
10089 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10091 (su): Add mappings for smax, smin, umax, umin.
10093 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10094 (FMAXMIN): Rename as...
10095 (FMAXMIN_UNS): ...This.
10097 (fmaxminv): Likewise.
10098 (fmaxmin): Likewise.
10100 (maxmin_uns_op): Likewise.
10102 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10104 * config/aarch64/arm_neon.h
10105 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10106 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10107 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10109 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10111 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10112 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10114 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10116 * config/aarch64/aarch64-simd.md
10117 (vcond<mode>_internal): Handle special cases for constant masks.
10118 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10119 (vcondu<mode><mode>): Likewise.
10120 (vcond<v_cmp_result><mode>): New.
10122 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10124 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10125 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10126 * config/aarch64/aarch64-simd-builtins.def
10127 (cmeq): Update to BUILTIN_VALLDI.
10132 * config/aarch64/arm_neon.h
10133 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10134 to builtins or C as appropriate.
10136 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10138 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10140 (cmhi): Rename to...
10142 * config/aarch64/aarch64-simd.md
10143 (simd_mode): Add SF.
10144 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10145 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10146 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10147 (cstore<mode>_neg): ...This.
10148 * config/aarch64/iterators.md
10150 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10151 (COMPARISONS): New.
10152 (UCOMPARISONS): Likewise.
10153 (optab): Add missing comparisons.
10159 (VCMP_S): Likewise.
10160 (VCMP_U): Likewise.
10161 (V_cmp_result): Add DF, SF modes.
10162 (v_cmp_result): Likewise.
10164 (vmtype): Likewise.
10165 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10167 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10169 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10170 define_insn to define_insn_and_split.
10171 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10172 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10173 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10174 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10175 (thumb2_negscc): Likewise.
10177 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10179 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10181 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10183 * config/arm/thumb2.md: Remove trailing whitespaces.
10185 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10187 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10188 Use gen_int_mode rather than GEN_INT.
10190 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10192 * value-prof.c (stream_in_histogram_value): Remove the strayed
10195 2013-04-30 Richard Biener <rguenther@suse.de>
10197 PR middle-end/57122
10198 * cfghooks.c (split_edge): Properly check for the loop latch edge.
10200 2013-04-30 Richard Biener <rguenther@suse.de>
10202 PR middle-end/57107
10203 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
10205 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
10207 PR rtl-optimization/56957
10208 PR rtl-optimization/57105
10209 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
10210 variable. Use just INSN_UID for determining whether an insn
10211 should be only disconnected from the insn stream.
10212 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
10214 2013-04-30 Jakub Jelinek <jakub@redhat.com>
10216 PR tree-optimization/57104
10217 * tsan.c (instrument_expr): Don't instrument accesses to
10218 DECL_HARD_REGISTER VAR_DECLs.
10220 2013-04-30 Richard Biener <rguenther@suse.de>
10222 * function.h (loops_for_fn): New inline function.
10223 (set_loops_for_fn): Likewise.
10224 * cfgloop.h (place_new_loop): Add struct function parameter.
10225 (get_loop): Likewise.
10226 (get_loops): Likewise.
10227 (number_of_loops): Likewise.
10228 (fel_next): Adjust.
10229 (fel_init): Likewise.
10230 * cfg.c (get_loop_copy): Adjust.
10231 * cfgloop.c (flow_loops_dump): Likewise.
10232 (record_loop_exits): Likewise.
10233 (verify_loop_structure): Likewise.
10234 * cfgloopanal.c (mark_irreducible_loops): Likewise.
10235 (estimate_reg_pressure_cost): Likewise.
10236 (mark_loop_exit_edges): Likewise.
10237 * cfgloopmanip.c (place_new_loop): Likewise.
10238 (add_loop): Likewise.
10239 (duplicate_loop): Likewise.
10240 * graph.c (draw_cfg_nodes): Likewise.
10241 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
10242 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
10243 (extract_affine_chrec): Likewise.
10244 (build_scop_iteration_domain): Likewise.
10245 * graphite.c (graphite_initialize): Likewise.
10246 * ira-build.c (create_loop_tree_nodes): Likewise.
10247 (more_one_region_p): Likewise.
10248 (rebuild_regno_allocno_maps): Likewise.
10249 (mark_loops_for_removal): Likewise.
10250 (mark_all_loops_for_removal): Likewise.
10251 (remove_unnecessary_regions): Likewise.
10252 (ira_build): Likewise.
10253 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
10254 * loop-init.c (fix_loop_structure): Likewise.
10255 (gate_rtl_move_loop_invariants): Likewise.
10256 (gate_rtl_unswitch): Likewise.
10257 (gate_rtl_unroll_and_peel_loops): Likewise.
10258 (rtl_doloop): Likewise.
10259 * lto-streamer-in.c (input_cfg): Likewise.
10260 * lto-streamer-out.c (output_cfg): Likewise.
10261 * modulo-sched.c (sms_schedule): Likewise.
10262 * predict.c (tree_estimate_probability): Likewise.
10263 (tree_estimate_probability_driver): Likewise.
10264 (estimate_loops): Likewise.
10265 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
10266 (move_sese_region_to_fn): Likewise.
10267 (debug_loop_num): Likewise.
10268 * tree-chrec.c (chrec_evaluate): Likewise.
10269 (hide_evolution_in_other_loops_than_loop): Likewise.
10270 (chrec_component_in_loop_num): Likewise.
10271 (reset_evolution_in_loop): Likewise.
10272 (evolution_function_is_invariant_rec_p): Likewise.
10273 * tree-if-conv.c (main_tree_if_conversion): Likewise.
10274 * tree-inline.c (copy_loops): Likewise.
10275 (copy_cfg_body): Likewise.
10276 (tree_function_versioning): Likewise.
10277 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
10278 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
10280 (add_to_evolution_1): Likewise.
10281 (scev_const_prop): Likewise.
10282 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
10283 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
10284 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
10285 (tree_ssa_lim_initialize): Likewise.
10286 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
10287 (verify_loop_closed_ssa): Likewise.
10288 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
10289 (tree_ssa_loop_im): Likewise.
10290 (tree_ssa_loop_unswitch): Likewise.
10291 (tree_vectorize): Likewise.
10292 (check_data_deps): Likewise.
10293 (tree_ssa_loop_ivcanon): Likewise.
10294 (tree_ssa_loop_bounds): Likewise.
10295 (tree_complete_unroll): Likewise.
10296 (tree_complete_unroll_inner): Likewise.
10297 (tree_parallelize_loops): Likewise.
10298 (tree_ssa_loop_prefetch): Likewise.
10299 (tree_ssa_loop_ivopts): Likewise.
10300 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
10301 * tree-vectorizer.c (vectorize_loops): Likewise.
10303 2013-04-29 Mike Frysinger <vapier@gentoo.org>
10305 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
10306 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
10307 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
10308 with EABI_LINK_SPEC.
10310 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10313 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
10316 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
10319 * lra-constraints.c (process_alt_operands): Discourage a bit more
10320 using memory for pseudos. Print cost dump for alternatives.
10321 Modify cost values for conflicts with early clobbers.
10322 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
10324 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10327 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
10329 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10331 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
10332 from/to S register.
10333 (movdi_aarch64): Support LDR/STR from/to D register.
10335 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10337 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
10338 or higher by default.
10340 2013-04-29 Richard Biener <rguenther@suse.de>
10342 PR middle-end/57075
10343 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
10344 even if not adding abnormal edges for calls that can make
10347 2013-04-29 Richard Biener <rguenther@suse.de>
10349 PR middle-end/57103
10350 * tree-cfg.c (move_stmt_op): Fix condition under which to update
10352 (move_stmt_r): Remove redundant checking.
10354 2013-04-29 Teresa Johnson <tejohnson@google.com>
10357 * basic-block.h (apply_scale): New function.
10358 (apply_probability): Use apply_scale.
10359 * gimple-streamer-in.c (input_bb): Ditto.
10360 * lto-streamer-in.c (input_cfg): Ditto.
10361 * lto-cgraph.c (merge_profile_summaries): Ditto.
10362 * tree-optimize.c (execute_fixup_cfg): Ditto.
10363 * tree-inline.c (copy_bb): Update comment to use apply_scale.
10364 (copy_edges_for_bb): Ditto.
10365 (copy_cfg_body): Ditto.
10367 2013-04-29 Tom de Vries <tom@codesourcery.com>
10369 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
10370 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
10371 (tail_merge_optimize): Handle current_loops == NULL.
10373 2013-04-26 Jeff Law <law@redhat.com>
10375 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
10376 (simplify_cond_using_ranges): Generalize code to simplify
10377 COND_EXPRs where one argument is a constant and the other
10378 is an SSA_NAME created by an integral type conversion.
10380 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10382 * config/arm/arm.md (store_minmaxsi): Use only when
10383 optimize_insn_for_size_p.
10385 2013-04-29 Christian Bruel <christian.bruel@st.com>
10388 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
10390 2013-04-29 Richard Biener <rguenther@suse.de>
10392 PR middle-end/57089
10393 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
10394 loop tree make sure to schedule a fixup for the child as well.
10395 (expand_omp_for_generic): Properly add loops.
10396 (expand_omp_for_static_nochunk): Likewise.
10397 (expand_omp_for_static_chunk): Likewise.
10398 (expand_omp_for): For the degenerate case fixup loops.
10399 (expand_omp_sections): Fix default bb placement in loops.
10400 (expand_omp_atomic_pipeline): Properly add loops.
10402 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10404 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
10406 2013-04-29 Tom de Vries <tom@codesourcery.com>
10408 * tree-ssa-tail-merge.c: Update header comment.
10410 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10412 * config/aarch64/arm_neon.h
10413 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
10414 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
10415 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
10416 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
10417 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
10418 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
10419 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
10420 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
10422 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10424 * config/aarch64/aarch64-simd.md
10425 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
10426 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
10427 fix_trunc, fixuns_trunc.
10428 (ftrunc<VDQF:mode>2): New.
10429 * config/aarch64/iterators.md (optab): Add fix, fixuns.
10430 (fix_trunc_optab): New.
10432 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10434 * config/aarch64/aarch64-builtins.c
10435 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
10436 iceilf, lround, iroundf.
10438 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10441 * config/i386/i386.h (enum ix86_tune_indices)
10442 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
10443 New, split from X86_TUNE_INTER_UNIT_MOVES.
10444 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
10445 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
10446 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
10447 (TARGET_INTER_UNIT_MOVES): Remove.
10448 * config/i386/i386.c (initial_ix86_tune_features): Update.
10449 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
10450 (ix86_expand_convert_uns_didf_sse): Use
10451 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10452 (ix86_expand_vector_init_one_nonzero): Ditto.
10453 (ix86_expand_vector_init_interleave): Ditto.
10454 (inline_secondary_memory_needed): Return true for moves from SSE class
10455 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
10456 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
10457 * config/i386/constraints.md (Yi, Ym): Depend on
10458 TARGET_INTER_UNIT_MOVES_TO_VEC.
10459 (Yj, Yn): New constraints.
10460 * config/i386/i386.md (*movdi_internal): Change constraints of
10461 operand 1 from Yi to Yj and from Ym to Yn.
10462 (*movsi_internal): Ditto.
10463 (*movdf_internal): Ditto.
10464 (*movsf_internal): Ditto.
10465 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10466 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10467 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10468 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10469 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10470 * config/i386/sse.md (movdi_to_sse): Ditto.
10471 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10472 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10473 TARGET_INTER_UNIT_MOVES.
10474 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10475 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10476 instead of TARGET_INTER_UNIT_MOVES.
10477 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10478 operand 1 from Yi to Yj and from Ym to Yn.
10480 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10482 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10483 (float_truncate_hi_): Likewise.
10484 (float_extend_lo_): Likewise.
10485 (float_truncate_lo_): Likewise.
10486 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10487 (aarch64_float_extend_lo_v2df): Likewise.
10488 (vec_unpacks_hi_v4sf): Likewise.
10489 (aarch64_float_truncate_lo_v2sf): Likewise.
10490 (aarch64_float_truncate_hi_v4sf): Likewise.
10491 (vec_pack_trunc_v2df): Likewise.
10492 (vec_pack_trunc_df): Likewise.
10494 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10496 * config/aarch64/aarch64-builtins.c
10497 (aarch64_fold_builtin): Fold float conversions.
10498 * config/aarch64/aarch64-simd-builtins.def
10499 (floatv2si, floatv4si, floatv2di): New.
10500 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10501 * config/aarch64/aarch64-simd.md
10502 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10503 * config/aarch64/iterators.md (FLOATUORS): New.
10504 (optab): Add float, floatuns.
10505 (su_optab): Likewise.
10507 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10509 * config/aarch64/aarch64-builtins.c
10510 (aarch64_builtin_vectorized_function): Use new names for
10512 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10513 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10514 (fcvtzu): Split as...
10515 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10516 (fcvtas): Split as...
10517 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10518 (fcvtau): Split as...
10519 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10520 (fcvtps): Split as...
10521 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10522 (fcvtpu): Split as...
10523 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10524 (fcvtms): Split as...
10525 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10526 (fcvtmu): Split as...
10527 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10528 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10529 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10530 (lfrintnusf, lfrintnudf): Likewise.
10531 * config/aarch64/aarch64-simd.md
10532 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10534 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10535 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10536 (fcvt_pattern): Likewise.
10538 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10540 * config/aarch64/aarch64-simd.md
10541 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10542 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10544 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10546 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10547 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10548 (vrnd<a,m,n,p>_f32): Implement using builtins.
10549 (vrnd<i,x><q>_f<32, 64>): New.
10551 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10553 * config/aarch64/aarch64-builtins.c
10554 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10555 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10556 (frintz): Rename to...
10558 (frintp): Rename to...
10560 (frintm): Rename to...
10562 (frinti): Rename to...
10563 (nearbyint): ...this.
10564 (frintx): Rename to...
10566 (frinta): Rename to...
10568 * config/aarch64/aarch64-simd.md
10569 (aarch64_frint<frint_suffix><mode>): Delete.
10570 (<frint_pattern><mode>2): Convert to insn.
10571 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10572 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10573 (frint_pattern): Likewise.
10574 (frint_suffix): Likewise.
10576 2013-04-29 Richard Biener <rguenther@suse.de>
10578 PR tree-optimization/57081
10579 * loop-init.c: Include tree-flow.h.
10580 (loop_optimizer_finalize): Free number of iteration estimates.
10581 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10583 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10585 PR tree-optimization/57083
10586 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10587 non-singleton shift count range, zero extend low_bound for uns case.
10589 * config/i386/predicates.md (general_vector_operand): New predicate.
10590 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10591 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10592 if they aren't nonimmediate operands. If their original values
10593 satisfy const_vector_equal_evenodd_p, don't shift them.
10594 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10595 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10596 if not nonimmediate_operand.
10597 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10598 instead of register_operand.
10599 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10601 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10603 * stor-layout.c (finalize_size_functions): Allocate a structure and
10604 reset cfun before dumping the functions.
10606 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10608 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10611 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10612 use xop_pmacsdqh if uns_p.
10613 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10614 the immediate rotate count.
10616 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10618 * rtl.h (struct rtx_def): Add comment for field jump.
10619 (LRA_SUBREG_P): New macro.
10620 * recog.c (register_operand): Check LRA_SUBREG_P.
10621 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10623 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10624 (lra_final_code_change): Skip subreg change for operators.
10625 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10626 if there are no operand changes.
10627 * lra-constraints.c (curr_insn_set): New.
10628 (match_reload): Set LRA_SUBREG_P.
10629 (emit_spill_move): Ditto.
10630 (check_and_process_move): Use curr_insn_set. Process only single
10631 set insns. Don't initialize sec_mem_p and change_p.
10632 (simplify_operand_subreg): Use LRA_SUBREG_P.
10633 (reg_in_class_p): New function.
10634 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10635 of #ifdef. Add code to remove cycling.
10636 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10637 non-null disp. Reload inner instead of disp when base and index
10638 are null. Try to put lo_sum into register.
10639 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10640 (check_and_process_move): Move code for move cost check to
10641 simple_move_p. Remove equiv_substitution.
10642 (simple_move_p): New function.
10643 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10644 curr_insn_set. Call check_and_process_move only for single set
10645 insns. Use the new function. Move call of check_and_process_move
10646 after operand equiv substitution and address process.
10648 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10651 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10652 with nonlocal goto receivers or returns twice calls, ignore
10653 unininitialized values from abnormal edges to nl goto receiver
10654 or returns twice call.
10656 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10658 PR tree-optimization/57051
10659 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10660 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10663 2013-04-26 Richard Biener <rguenther@suse.de>
10665 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10666 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10667 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10668 (fixup_loop_arrays_after_move): New function.
10669 (move_sese_region_to_fn): Properly outline the loop tree parts
10670 of the SESE region.
10672 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10674 * config/i386/i386.md (type, unit): Fix long lines.
10676 2013-04-26 Richard Biener <rguenther@suse.de>
10678 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10679 (lto-streamer-out.o): Likewise.
10680 * cfgloop.c (init_loops_structure): Export, add struct function
10681 argument and adjust.
10682 (flow_loops_find): Adjust.
10683 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10684 (init_loops_structure): Declare.
10685 * lto-streamer-in.c: Include cfgloop.h.
10686 (input_cfg): Input the loop tree.
10687 * lto-streamer-out.c: Include cfgloop.h.
10688 (output_cfg): Output the loop tree.
10689 (output_struct_function_base): Do not drop PROP_loops.
10691 2013-03-26 Richard Biener <rguenther@suse.de>
10693 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10694 (pass_build_cfg): Provide PROP_loops.
10695 (move_sese_region_to_fn): Remove loops that are outlined into fn
10697 * tree-inline.c: Include cfgloop.h.
10698 (initialize_cfun): Do not drop PROP_loops.
10699 (copy_loops): New function.
10700 (copy_cfg_body): Copy loop structure.
10701 (tree_function_versioning): Initialize destination loop tree.
10702 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10703 (pass_parallelize_loops): Do IL verification.
10704 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10705 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10706 the CFG make sure we fixup loops as well.
10707 * tree-ssa-tail-merge.c: Include cfgloop.h.
10708 (replace_block_by): When merging loop latches mark loops for fixup.
10709 * lto-streamer-out.c (output_struct_function_base): Drop
10710 PROP_loops for now.
10711 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10712 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10713 * ipa-split.c: Include cfgloop.h.
10714 (split_function): Add the new return block to the loop tree root.
10715 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10716 whether we have removed the forwarder block.
10717 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10718 * cfgloop.h (place_new_loop): Declare.
10719 * cfgloopmanip.c (place_new_loop): Export.
10720 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10721 (tree-switch-conversion.o): Likewise.
10722 (tree-complex.o): Likewise.
10723 (tree-inline.o): Likewise.
10724 (tree-ssa-tailmerge.o): Likewise.
10725 (ipa-split.o): Likewise.
10726 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10727 (tree-ssa-copy.o): Likewise.
10728 * tree-switch-conversion.c: Include cfgloop.h
10729 (process_switch): If we emit a bit-test cascade, schedule loops
10731 * tree-complex.c: Include cfgloop.h.
10732 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10733 * asan.c: Include cfgloop.h.
10734 (create_cond_insert_point): Properly add new basic-blocks to
10735 loops, schedule loop fixup.
10736 * cfgloop.c (verify_loop_structure): Check that looks are not
10738 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10740 (expand_omp_for_generic): Likewise.
10741 (expand_omp_sections): Likewise.
10742 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10743 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10744 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10745 is initialized, not when loops are present.
10746 * tree-parloops.c (parallelize_loops): Remove checking here.
10747 * passes.c (init_optimization_passes): Schedule a copy-propagation
10748 pass before complete unrolling of inner loops.
10750 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10752 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10753 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10754 (colorize_init): Add argument to _WIN32 version.
10755 * toplev.c: Include diagnostic-color.h.
10756 (process_options): Default to -fdiagnostics-color=auto if
10757 GCC_COLORS env var is in the environment.
10758 * common.opt (fdiagnostics-color=): Add Var and Init.
10759 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10760 env var is in the environment, the default is auto rather than never.
10762 * diagnostic.h (file_name_as_prefix): Add context argument.
10763 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10764 the string as locus.
10765 * langhooks.c (lhd_print_error_function): Adjust caller.
10767 2013-04-25 Lawrence Crowl <crowl@google.com>
10769 * var-tracking.c (shared_hash_def::htab):
10770 Change type to hash_table. Update dependent calls and types.
10772 2013-04-25 Lawrence Crowl <crowl@google.com>
10774 * Makefile.in: Update as needed below.
10776 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10777 Move declaration to after the type's method definitons.
10779 * attribs.c (htab_t scoped_attributes::attribute_hash):
10780 Change type to hash_table. Update dependent calls and types.
10782 * bitmap.c (htab_t bitmap_desc_hash):
10783 Change type to hash_table. Update dependent calls and types.
10785 * cselib.c (htab_t cselib_hash_table):
10786 Change type to hash_table. Update dependent calls and types.
10788 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10789 (hash_string_slot_node): Move implementation into lto-streamer.h
10790 struct string_slot_hasher.
10791 (eq_string_slot_node): Likewise.
10793 * data-streamer-out.c: Update output_block::string_hash_table
10794 dependent calls and types.
10796 * dwarf2cfi.c (htab_t trace_index):
10797 Change type to hash_table. Update dependent calls and types.
10799 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10800 Change type to hash_table. Update dependent calls and types.
10801 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10802 (htab_t optimize_external_refs::map): Likewise.
10803 (htab_t output_comp_unit::extern_map): Likewise.
10804 (htab_t output_comdat_type_unit::extern_map): Likewise.
10805 (htab_t output_macinfo::macinfo_htab): Likewise.
10806 (htab_t optimize_location_lists::htab): Likewise.
10807 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10809 * except.c (htab_t ehspec_hash_type):
10810 Change type to hash_table. Update dependent calls and types.
10811 (assign_filter_values::ttypes): Likewise.
10812 (assign_filter_values::ehspec): Likewise.
10813 (sjlj_assign_call_site_values::ar_hash): Likewise.
10814 (convert_to_eh_region_ranges::ar_hash): Likewise.
10816 * gcse.c (htab_t pre_ldst_table):
10817 Change type to hash_table. Update dependent calls and types.
10819 * ggc-common.c (htab_t saving_htab):
10820 Change type to hash_table. Update dependent calls and types.
10821 (htab_t loc_hash): Likewise.
10822 (htab_t ptr_hash): Likewise.
10823 (call_count): Rename ggc_call_count.
10824 (call_alloc): Rename ggc_call_alloc.
10825 (loc_descriptor): Rename make_loc_descriptor.
10826 (add_statistics): Rename ggc_add_statistics.
10828 * ggc-common.c (saving_htab):
10829 Change type to hash_table. Update dependent calls and types.
10831 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10832 (push_gimplify_context): Likewise.
10833 (pop_gimplify_context): Likewise.
10834 (struct gimple_temp_hash_elt): Added.
10835 (struct gimplify_hasher): Likewise.
10836 (struct gimplify_ctx.temp_htab):
10837 Change type to hash_table. Update dependent calls and types.
10839 * gimple-fold.c: Include gimplify-ctx.h.
10841 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10842 Change type to hash_table. Update dependent calls and types.
10843 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10844 avoid potential global name collision.
10846 * gimplify.c: Include gimplify-ctx.h.
10847 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10848 (htab_t gimplify_ctx::temp_htab):
10849 Update dependent calls and types for new type hash_table.
10850 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10851 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10853 * gimplify-ctx.h: New.
10854 (struct gimple_temp_hash_elt): Move from gimplify.c.
10855 (class gimplify_hasher): New.
10856 (struct gimplify_ctx): Move from gimple.h.
10857 (htab_t gimplify_ctx::temp_htab):
10858 Change type to hash_table. Update dependent calls and types.
10860 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10861 (htab_t ivs_params::newivs_index):
10862 Change type to hash_table. Update dependent calls and types.
10863 (htab_t ivs_params::params_index): Likewise.
10864 (htab_t print_generated_program::params_index): Likewise.
10865 (htab_t gloog::newivs_index): Likewise.
10866 (htab_t gloog::params_index): Likewise.
10868 * graphite.c: Include graphite-htab.h.
10869 4htab_t graphite_transform_loops::bb_pbb_mapping):
10870 Change type to hash_table. Update dependent calls and types.
10872 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10873 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10874 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10876 * graphite-dependences.c: Include graphite-htab.h.
10877 (loop_is_parallel_p): Change hash table type of parameter.
10879 * graphite-htab.h: New.
10880 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10881 (extern find_pbb_via_hash): Move from graphite-poly.h.
10882 (extern loop_is_parallel_p): Move from graphite-poly.h.
10883 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10885 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10886 (extern loop_is_parallel_p): Move to graphite-htab.h.
10887 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10889 * haifa-sched.c (htab_t delay_htab):
10890 Change type to hash_table. Update dependent calls and types.
10891 (htab_t delay_htab_i2): Likewise.
10893 * ira-color.c (htab_t allocno_hard_regs_htab):
10894 Change type to hash_table. Update dependent calls and types.
10896 * ira-costs.c (htab_t cost_classes_htab):
10897 Change type to hash_table. Update dependent calls and types.
10899 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10900 Change type to hash_table. Update dependent calls and types.
10902 * loop-iv.c (htab_t bivs):
10903 Change type to hash_table. Update dependent calls and types.
10905 * loop-unroll.c (htab_t opt_info::insns_to_split):
10906 Change type to hash_table. Update dependent calls and types.
10907 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10909 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10910 (struct string_slot_hasher): New.
10911 (htab_t output_block::string_hash_table):
10912 Change type to hash_table. Update dependent calls and types.
10914 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10915 (htab_t file_name_hash_table):
10916 Change type to hash_table. Update dependent calls and types.
10918 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10921 * lto-streamer.c (htab_t tree_htab):
10922 Change type to hash_table. Update dependent calls and types.
10924 * omp-low.c: Include gimplify-ctx.h.
10926 * passes.c (htab_t name_to_pass_map):
10927 Change type to hash_table. Update dependent calls and types.
10928 (pass_traverse): Rename to passes_pass_traverse.
10930 * plugin.c (htab_t event_tab):
10931 Change type to hash_table. Update dependent calls and types.
10933 * postreload-gcse.c (htab_t expr_table):
10934 Change type to hash_table. Update dependent calls and types.
10935 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10937 * sese.c (debug_rename_map_1): Make extern.
10938 (htab_t copy_bb_and_scalar_dependences::rename_map):
10939 Change type to hash_table. Update dependent calls and types.
10941 * sese.h (extern debug_rename_map): Move to .c file.
10943 * store-motion.c (htab_t store_motion_mems_table):
10944 Change type to hash_table. Update dependent calls and types.
10946 * trans-mem.c (htab_t tm_new_mem_hash):
10947 Change type to hash_table. Update dependent calls and types.
10949 * tree-browser.c (htab_t TB_up_ht):
10950 Change type to hash_table. Update dependent calls and types.
10952 * tree-cfg.c (htab_t discriminator_per_locus):
10953 Change type to hash_table. Update dependent calls and types.
10955 * tree-complex.c: Include tree-hasher.h
10956 (htab_t complex_variable_components):
10957 Change type to hash_table. Update dependent calls and types.
10959 * tree-eh.c (htab_t finally_tree):
10960 Change type to hash_table. Update dependent calls and types.
10962 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10963 struct int_tree_hasher.
10964 (extern int_tree_map_eq): Likewise.
10965 (uid_decl_map_hash): Removed.
10966 (extern decl_tree_map_eq): Likewise.
10968 * tree-hasher.h: New.
10969 (struct int_tree_hasher): New.
10970 (typedef int_tree_htab_type): New.
10972 * tree-inline.c: Include gimplify-ctx.h.
10974 * tree-mudflap.c: Include gimplify-ctx.h.
10976 * tree-parloops.c: Include tree-hasher.h.
10977 (htab_t eliminate_local_variables_stmt::decl_address):
10978 Change type to hash_table. Update dependent calls and types.
10979 (htab_t separate_decls_in_region::decl_copies): Likewise.
10981 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10982 Change type to hash_table. Update dependent calls and types.
10984 * tree-sra.c (candidates):
10985 Change type to hash_table. Update dependent calls and types.
10987 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10989 (int_tree_map_hash): Likewise.
10991 * tree-ssa-dom.c (htab_t avail_exprs):
10992 Change type to hash_table. Update dependent calls and types.
10994 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10995 Change type to hash_table. Update dependent calls and types.
10997 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10998 Change type to hash_table. Update dependent calls and types.
11000 * tree-ssa-phiopt.c (seen_ssa_names):
11001 Change type to hash_table. Update dependent calls and types.
11003 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11004 Change type to hash_table. Update dependent calls and types.
11006 * tree-ssa-uncprop.c (equiv):
11007 Change type to hash_table. Update dependent calls and types.
11009 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11011 PR rtl-optimization/57003
11012 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11013 call note_stores with kill_clobbered_value callback again after
11014 killing regs_invalidated_by_call.
11016 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11018 * config/aarch64/aarch64-simd.md
11019 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11020 (aarch64_simd_bsl<mode>): Likewise.
11021 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11023 2013-04-25 Marek Polacek <polacek@redhat.com>
11025 PR tree-optimization/57066
11026 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11028 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11030 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11032 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11034 * config/aarch64/aarch64-builtins.c
11035 (aarch64_fold_builtin): New.
11036 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11037 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11038 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11039 * config/aarch64/arm_neon.h
11040 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11042 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11043 Tejas Belagod <tejas.belagod@arm.com>
11045 * config/aarch64/aarch64-builtins.c
11046 (aarch64_gimple_fold_builtin): New.
11047 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11048 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11049 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11050 (addvv4sf): Update.
11051 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11053 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11055 * config/aarch64/aarch64.md
11056 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11058 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11060 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11061 (*ngcsi_uxtw): New pattern.
11063 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11064 Julian Brown <julian@codesourcery.com>
11066 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11067 (TB_DREG): Add T_V4HF.
11068 (v4hf_UP): New macro.
11069 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11070 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11071 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11073 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11074 (arm_vector_mode_supported_p): Handle V4HF.
11075 (arm_mangle_map): Handle V4HFmode.
11076 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11077 * config/arm/arm_neon_builtins.def: Add entries for
11078 vcvtv4hfv4sf, vcvtv4sfv4hf.
11079 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11080 (neon_vcvtv4hfv4sf): Likewise.
11081 * config/arm/neon-gen.ml: Handle half-precision floating point
11083 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11084 * config/arm/arm_neon.h: Regenerate.
11085 * config/arm/neon.ml (type elts): Add F16.
11086 (type vectype): Add T_float16x4, T_floatHF.
11087 (type vecmode): Add V4HF.
11088 (type features): Add Requires_FP_bit feature.
11089 (elt_width): Handle F16.
11090 (elt_class): Likewise.
11091 (elt_of_class_width): Likewise.
11092 (mode_of_elt): Refactor.
11093 (type_for_elt): Handle F16, fix error messages.
11094 (vectype_size): Handle T_float16x4.
11095 (vcvt_sh): New function.
11096 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11097 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11098 (string_of_mode): Handle V4HF.
11099 * doc/arm-neon-intrinsics.texi: Regenerate.
11101 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11103 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11104 format specifier in 'X' case.
11106 2013-04-25 Alan Modra <amodra@gmail.com>
11109 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11110 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11111 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11112 Repeat for many other rotate/shift and mask patterns using subregs.
11113 Name lshiftrt insns.
11114 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11115 on WORDS_BIG_ENDIAN.
11117 2013-04-25 Alan Modra <amodra@gmail.com>
11119 * config.gcc: Support little-endian powerpc-linux targets.
11120 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11121 (LINK_OS_LINUX_SPEC): Define.
11122 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11123 Preserve MASK_LITTLE_ENDIAN.
11124 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11125 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11126 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11127 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11128 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11129 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11130 Correct fp word order for little-endian. Don't shift toc entries
11131 smaller than a word for little-endian.
11132 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11133 (bswapdi2 splits): Correct low-part subreg for little-endian.
11134 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11135 low/high where such is correct only for be.
11136 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11137 little-endian for -mcall-aixdesc.
11139 2013-04-25 Alan Modra <amodra@gmail.com>
11141 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11142 replace_equiv_address_nv.
11144 2013-04-25 Alan Modra <amodra@gmail.com>
11146 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11148 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11151 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11152 * rtl.h (struct rtx_def): ...
11154 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11156 PR rtl-optimizations/57046
11157 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11158 for multi-reg splits.
11160 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11162 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11164 2013-04-24 Sterling Augustine <saugustine@google.com>
11166 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11167 (comp_dir_string, debug_str_dwo_section): New.
11168 (DEBUG_STR_DWO_SECTION): Rename to ...
11169 (DEBUG_DWO_STR_SECTION): ... this.
11170 (DEBUG_NORM_STR_SECTION): Delete.
11171 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11172 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11173 (find_AT_string): Move most logic to ...
11174 (find_AT_string_in_table): ... here. New.
11175 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11176 add_skeleton_AT_string. Delete logic.
11177 (output_skeleton_debug_sections): Remove call to
11178 add_top_level_skeleton_die_attrs.
11179 (add_comp_dir_attribute): Move logic to comp_dir_string.
11180 (dwarf2out_init): Initialize debug_str_dwo_section.
11181 (output_indirect_string): Call find_string_form.
11182 (output_indirect_strings): Rewrite.
11183 (prune_unused_types): Empty skeleton_debug_str_hash.
11184 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11185 (dwarf2out_finish): Call output_indirect_strings.
11187 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11189 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11191 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11193 * rtl.h (struct rtx_def): Add comment for field jump.
11194 (LRA_SUBREG_P): New macro.
11195 * recog.c (register_operand): Check LRA_SUBREG_P.
11196 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11198 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11199 (lra_final_code_change): Skip subreg change for operators.
11200 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11201 if there are no operand changes.
11202 * lra-constraints.c (curr_insn_set): New.
11203 (match_reload): Set LRA_SUBREG_P.
11204 (emit_spill_move): Ditto.
11205 (check_and_process_move): Use curr_insn_set. Process only single
11206 set insns. Don't initialize sec_mem_p and change_p.
11207 (simplify_operand_subreg): Use LRA_SUBREG_P.
11208 (reg_in_class_p): New function.
11209 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11210 of #ifdef. Add code to remove cycling.
11211 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11212 non-null disp. Reload inner instead of disp when base and index
11213 are null. Try to put lo_sum into register.
11214 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11215 (check_and_process_move): Move code for move cost check to
11216 simple_move_p. Remove equiv_substitution.
11217 (simple_move_p): New function.
11218 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11219 curr_insn_set. Call check_and_process_move only for single set
11220 insns. Use the new function. Move call of check_and_process_move
11221 after operand equiv substitution and address process.
11223 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
11225 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
11226 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
11227 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
11229 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11231 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
11233 2013-04-24 Marek Polacek <polacek@redhat.com>
11235 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
11236 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
11237 (select_loops_exit_conditions): Likewise.
11238 (number_of_iterations_for_all_loops): Likewise.
11239 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
11240 (scev_analysis): Likewise.
11242 2013-04-02 Catherine Moore <clm@codesourcery.com>
11243 Chao-ying Fu <fu@mips.com>
11245 * config/mips/micromips.md (jraddiusp): New pattern.
11246 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
11247 instruction if possible.
11249 2013-04-24 Alan Modra <amodra@gmail.com>
11251 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
11253 2013-04-24 Julian Brown <julian@codesourcery.com>
11254 Chung-Lin Tang <cltang@codesourcery.com>
11256 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
11257 dependency behavior in enumeration type DIE generation. Add TODO note
11258 to comments about future DW_FORM_sdata/udata re-work of related code.
11260 2013-04-23 Lawrence Crowl <crowl@google.com>
11262 * Makefile.in: Update as needed below.
11264 * hash-table.h (class hash_table):
11265 Correct many methods with parameter types compare_type to the correct
11266 value_type. (Correct code was unlikely to notice the change.)
11267 (hash_table::elements_with_deleted) New.
11268 (class hashtable::iterator): New.
11269 (hashtable::begin()): New.
11270 (hashtable::end()): New.
11271 (FOR_EACH_HASH_TABLE_ELEMENT): New.
11273 * statistics.c (statistics_hashes):
11274 Change type to hash_table. Update dependent calls and types.
11276 * tree-into-ssa.c (var_infos):
11277 Change type to hash_table. Update dependent calls and types.
11279 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
11280 Change type to hash_table. Update dependent calls and types.
11282 * tree-ssa-loop-im.c (struct mem_ref.refs):
11283 Change type to hash_table. Update dependent calls and types.
11285 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
11286 Change type to hash_table. Update dependent calls and types.
11288 * tree-ssa-sccvn.c (vn_tables_s::nary):
11289 Change type to hash_table. Update dependent calls and types.
11290 (vn_tables_s::phis): Likewise.
11291 (vn_tables_s::references): Likewise.
11293 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
11294 (vn_reference_eq): Update parameter and return types.
11296 * tree-ssa-structalias.c (pointer_equiv_class_table):
11297 Change type to hash_table. Update dependent calls and types.
11298 (location_equiv_class_table): Likewise.
11300 * tree-vect-data-refs.c: Consequential changes for making
11301 peeling a hash_table.
11303 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
11304 (destroy_loop_vec_info): Dependent hash_table update.
11306 * tree-vectorizer.h (peeling_htab):
11307 Change type to hash_table. Update dependent calls and types.
11309 2013-04-23 Shiva Chen <shiva0217@gmail.com>
11311 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
11312 to check the register content is equal or not.
11313 * lra-constraints.c (match_reload): Use lra_assign_reg_val
11314 to assign register content record.
11315 * lra-eliminations.c (update_reg_eliminate): Use
11316 lra_update_reg_val_offset to update register content offset.
11317 * lra-int.h (struct lra_reg): Add offset member.
11318 (lra_reg_val_equal_p): New static inline function.
11319 (lra_update_reg_val_offset): New static inline function.
11320 (lra_assign_reg_val): New static inline function.
11321 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
11322 to assign register content record.
11323 (initialize_lra_reg_info_element): Initial offset to zero.
11325 2013-04-23 Catherine Moore <clm@codesourcery.com>
11327 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
11328 operands. Record compression.
11330 2013-04-23 Xinliang David Li <davidxl@google.com>
11332 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
11334 2013-04-23 Richard Biener <rguenther@suse.de>
11336 PR middle-end/57036
11337 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
11338 parameter, only add abnormal goto edges from the copied body
11339 if the call could perform abnormal gotos.
11340 (copy_cfg_body): Adjust.
11342 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
11344 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
11346 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
11348 * coretypes.h (gimple_stmt_iterator): Add struct to make
11351 2013-04-23 Richard Biener <rguenther@suse.de>
11353 PR tree-optimization/57026
11354 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
11355 from SSA names occuring in abnormal PHI nodes.
11357 2013-04-22 Andi Kleen <ak@linux.intel.com>
11359 * lto/lto.c (print_lto_report_1): Fix LTO report names.
11361 2013-04-22 Andi Kleen <ak@linux.intel.com>
11363 * lto/lto.c (print_lto_report_1): Declare early.
11364 (read_cgraph_and_symbols): Call print_lto_report_1 early.
11366 2013-04-22 Andi Kleen <ak@linux.intel.com>
11368 * common.opt (-flto-report-wpa): Add.
11369 * doc/invoke.texi (-flto-report-wpa): Add.
11370 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
11373 2013-04-22 Xinliang David Li <davidxl@google.com>
11375 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
11376 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
11377 * Makefile.in: New dependency
11379 David Daney <ddaney.cavm@gmail.com>
11381 * configure.ac (gcc_cv_as_micromips_support): Use the
11382 --fatal-warnings option.
11383 * configure: Regenerate.
11385 2013-04-22 Marek Polacek <polacek@redhat.com>
11388 * tsan.c (instrument_expr): Don't instrument expression
11389 in case its size is zero.
11391 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
11395 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
11397 * config/alpha/alpha.c (TARGET_LRA_P): New define.
11399 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11401 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
11402 (gimple_stmt_iterator): New typedef.
11403 * gimple.h (gimple_stmt_iterator): Rename to...
11404 (gimple_stmt_iterator_d): ... This.
11405 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
11406 trees be valid for GIMPLE and GENERIC.
11407 (TARGET_GIMPLE_FOLD_BUILTIN): New.
11408 * gimple-fold.c (gimple_fold_call): Call target hook
11409 gimple_fold_builtin.
11410 * hooks.c (hook_bool_gsiptr_false): New.
11411 * hooks.h (hook_bool_gsiptr_false): New.
11412 * target.def (fold_stmt): New.
11413 * doc/tm.texi: Regenerate.
11415 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
11418 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
11419 a set sp if no stack realignment.
11421 2013-04-22 Nick Clifton <nickc@redhat.com>
11423 * config.gcc (tilegx-linux): Extend extra_objs rather than
11425 (tilepro-linux): Likewise.
11427 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11429 * config/aarch64/aarch64-builtins.c
11431 (CF0, CF1, CF2, CF3, CF4, CF10): New.
11432 (VAR<1-12>): Add MAP parameter.
11433 (BUILTIN_*): Likewise.
11434 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
11435 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
11436 (aarch64_ushl_n<mode>): Likewise.
11437 (aarch64_sshr_n<mode>): Likewise.
11438 (aarch64_ushr_n<mode>): Likewise.
11439 (aarch64_<maxmin><mode>): Likewise.
11440 (aarch64_sqrt<mode>): Likewise.
11441 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
11442 (vshr<q>_n_*): Likewise.
11444 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11446 * config/aarch64/aarch64-builtins.c
11447 (aarch64_simd_builtin_type_mode): Handle SF types.
11449 (BUILTIN_GPF): Define.
11450 (aarch64_init_simd_builtins): Handle SF types.
11451 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
11452 (frecps): Likewise.
11453 (frecpx): Likewise.
11454 * config/aarch64/aarch64-simd.md
11455 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
11456 (aarch64_frecpe<mode>): New.
11457 (aarch64_frecps<mode>): Likewise.
11458 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
11459 (v8type): Add frecp<esx>.
11460 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
11461 (aarch64_frecps<mode>): Likewise.
11462 * config/aarch64/iterators.md (FRECP): New.
11463 (frecp_suffix): Likewise.
11464 * config/aarch64/arm_neon.h
11465 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11467 2013-04-22 Christian Bruel <christian.bruel@st.com>
11470 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11471 (REG_CLASS_NAMES): Idem.
11472 (REG_CLASS_CONTENTS): Idem.
11473 (REGCLASS_HAS_FP_REG): Idem.
11474 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11475 (sh_conditional_register_usage): Idem.
11477 2013-04-21 Jeff Law <law@redhat.com>
11479 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11480 (ssa_forward_propagate_and_combine): Use it.
11482 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11484 * lra.c: Update the flow chart diagram.
11486 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11488 PR rtl-optimization/56847
11489 * lra-constraints.c (process_alt_operands): Discourage alternative
11490 with non-matche doffsettable memory constraint fro memory with
11493 2013-04-19 Richard Biener <rguenther@suse.de>
11495 PR tree-optimization/56982
11496 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11498 * gimplify.c (gimplify_call_expr): Notice special calls.
11499 (gimplify_modify_expr): Likewise.
11500 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11501 abnormal control flow receivers.
11502 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11503 in the same way as cfun->has_nonlocal_labels.
11504 (gimple_purge_dead_abnormal_call_edges): Likewise.
11505 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11506 receivers start a basic-block.
11508 2013-04-19 Richard Biener <rguenther@suse.de>
11510 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11512 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11513 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11514 (SLP_TREE_LOAD_PERMUTATION): Add.
11515 (vect_transform_slp_perm_load): Adjust prototype.
11516 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11517 (vect_free_slp_instance): Likewise.
11518 (vect_create_new_slp_node): Likewise.
11519 (vect_supported_slp_permutation_p): Remove.
11520 (vect_slp_rearrange_stmts): Adjust.
11521 (vect_supported_load_permutation_p): Likewise. Inline
11522 vect_supported_slp_permutation_p here.
11523 (vect_analyze_slp_instance): Compute load permutations per
11524 slp node instead of per instance.
11525 (vect_get_slp_defs): Adjust.
11526 (vect_transform_slp_perm_load): Likewise.
11527 (vect_schedule_slp_instance): Remove redundant code.
11528 (vect_schedule_slp): Remove hack for PR56270, add it ...
11529 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11530 CSE loads for SLP. Adjust.
11532 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11534 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11535 spelling in two comments.
11537 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11540 * config/arm/arm.c (load_multiple_sequence): Require SP
11541 as base register for loads if SP is in the register list.
11543 2013-04-19 Martin Jambor <mjambor@suse.cz>
11545 PR tree-optimization/56718
11546 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11547 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11548 and made public. Adjusted all callers.
11549 (ipa_intraprocedural_devirtualization): New function.
11550 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11551 (ipa_intraprocedural_devirtualization): Likewise.
11552 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11554 2013-04-19 Richard Biener <rguenther@suse.de>
11556 PR tree-optimization/57000
11557 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11559 2013-04-19 Terry Guo <terry.guo@arm.com>
11561 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11563 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11564 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11565 (cortex_m4_fmacs): Use new reservations.
11566 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11568 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11570 PR rtl-optimization/56999
11571 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11573 (lra_coalesce): Remove split_origin_bitmap and related code.
11574 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11575 ranges if necessary.
11577 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11579 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11581 (ix86_expand_call): Remove clobbered_registers array and use
11582 x86_64_ms_sysv_extra_clobbered_registers instead.
11583 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11585 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11587 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11588 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11589 (*call_value_rex64_ms_sysv): Ditto.
11591 2013-04-18 Cary Coutant <ccoutant@google.com>
11593 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11594 parent instead of die_mark.
11596 2013-04-18 Diego Novillo <dnovillo@google.com>
11598 * gimple.c (create_gimple_tmp): New.
11599 (get_expr_type): New.
11600 (build_assign): New.
11601 (build_type_cast): New.
11602 * gimple.h (enum ssa_mode): Define.
11603 (gimple_seq_set_location): New.
11604 * asan.c (build_check_stmt): Change some gimple_build_* calls
11605 to use build_assign and build_type_cast.
11607 2013-04-18 Richard Biener <rguenther@suse.de>
11609 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11610 handle negative step. Remove redundant checks.
11611 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11612 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11613 for negative step and grouped loads fail to vectorize.
11615 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11617 * emit-rtl.c (reset_insn_used_flags): New function.
11618 (reset_all_used_flags): Use it.
11619 (verify_insn_sharing): New function.
11620 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11622 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11624 PR tree-optimization/56984
11625 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11626 and (x >> M) >= N don't register any assertion if N << M is the
11629 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11631 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11632 do not delete_insn insns that have not yet been emitted, only
11633 unlink them with remove_insn.
11634 * df-scan.c (df_insn_delete): Revert r197492.
11636 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11638 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11639 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11641 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11643 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11644 define_insn_and_split.
11645 (and_scc,ior_scc,negscc): Likewise.
11646 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11648 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11650 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11651 can benefit from using a sequence of LDRD instructions in epilogue
11652 instead of a single LDM instruction.
11654 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11657 * doc/extend.texi: Fix typo.
11659 2013-04-17 Richard Biener <rguenther@suse.de>
11661 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11662 (vect_build_slp_tree): ... here.
11663 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11664 match. Remove special-casing of mismatched complex loads.
11665 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11666 re-try the match with swapped commutative operands.
11667 (vect_supported_load_permutation_p): Remove special-casing of
11668 mismatched complex loads.
11669 (vect_analyze_slp_instance): Adjust.
11671 2013-04-17 Richard Biener <rguenther@suse.de>
11673 PR rtl-optimization/56921
11674 * cfgloop.h (struct loop): Add simple_loop_desc member.
11675 (struct niter_desc): Mark with GTY(()).
11676 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11677 * loop-iv.c (get_simple_loop_desc): Likewise.
11678 (free_simple_loop_desc): Likewise.
11681 2013-04-16 Richard Biener <rguenther@suse.de>
11683 PR rtl-optimization/56921
11684 * loop-init.c (pass_rtl_move_loop_invariants): Add
11685 TODO_do_not_ggc_collect to todo_flags_finish.
11686 (pass_rtl_unswitch): Same.
11687 (pass_rtl_unroll_and_peel_loops): Same.
11688 (pass_rtl_doloop): Same.
11690 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11692 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11693 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11694 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11696 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11697 * tree-streamer.c (record_common_node): Adjust reference in comment.
11699 2013-04-17 Terry Guo <terry.guo@arm.com>
11701 * config/arm/cortex-m4.md: Add a new bypass.
11703 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11705 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11707 (*subs_<optab><mode>_multp2): New pattern.
11708 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11709 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11711 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11713 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11714 (*subs_mul_imm_<mode>): New pattern.
11716 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11719 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11720 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11721 (vsx_movti_32bit): Change j->wa to O->wa.
11723 2013-04-16 Richard Biener <rguenther@suse.de>
11725 PR rtl-optimization/56921
11726 * loop-init.c (pass_rtl_move_loop_invariants): Add
11727 TODO_do_not_ggc_collect to todo_flags_finish.
11728 (pass_rtl_unswitch): Same.
11729 (pass_rtl_unroll_and_peel_loops): Same.
11730 (pass_rtl_doloop): Same.
11732 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11734 * config/arm/arm.c (emit_multi_reg_push): New declaration
11735 for an existing function.
11736 (arm_emit_strd_push): New function.
11737 (arm_expand_prologue): Used here.
11738 (arm_emit_ldrd_pop): New function.
11739 (arm_expand_epilogue): Used here.
11740 (arm_get_frame_offsets): Update condition.
11741 (arm_emit_multi_reg_pop): Add a special case for load of a single
11742 register with writeback.
11744 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11746 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11749 2013-04-16 Richard Biener <rguenther@suse.de>
11751 PR tree-optimization/56756
11752 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11753 (first_mem_ref_loc): New.
11754 (execute_sm): Place the load temporarily before a previous
11755 access instead of in the latch edge to ensure its SSA dependencies
11756 are defined at points dominating the load.
11758 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11760 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11761 correct fix by moving header and footer insn to the footer of
11762 the merged basic block. Clear BB_END of the merged-away block.
11764 PR middle-end/43631
11765 * emit-rtl.c (make_note_raw): New function.
11766 (link_insn_into_chain): New static inline function.
11767 (add_insn): Use it.
11768 (add_insn_before, add_insn_after): Factor insn chain linking code...
11769 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11770 using link_insn_into_chain.
11771 (note_outside_basic_block_p): New helper function for emit_note_after
11772 and emit_note_before.
11773 (emit_note_after): Use nobb variant of add_insn_after if the note
11774 should not be contained in a basic block.
11775 (emit_note_before): Use nobb variant of add_insn_before if the note
11776 should not be contained in a basic block.
11777 (emit_note_copy): Use make_note_raw.
11778 (emit_note): Likewise.
11779 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11780 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11781 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11782 the moved barrier the tail of the basic block it follows.
11783 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11785 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11787 PR tree-optimization/56962
11788 * gimple-ssa-strength-reduction.c (record_increment): Only set
11789 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11790 either rhs1 or rhs2 is equal to c->base_expr.
11792 2013-04-15 Richard Biener <rguenther@suse.de>
11794 PR tree-optimization/56933
11795 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11797 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11798 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11799 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11800 dependence check ...
11801 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11803 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11804 GROUP_READ_WRITE_DEPENDENCE.
11806 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11808 * emit-rtl.c (reset_all_used_flags): New function.
11809 (verify_rtl_sharing): Call reset_all_used_flags before and after
11810 performing the checks.
11812 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11814 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11815 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11816 * config/arm/constraints.md (De): New constraint.
11817 * config/arm/neon.md (anddi3_neon): Delete.
11818 (neon_vand<mode>): Expand to standard anddi3 pattern.
11819 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11820 Move earlier in the file.
11821 (neon_inv_logic_op2): Likewise.
11822 (arm_anddi_operand_neon): New predicate.
11824 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11826 * configure.ac (gcc_cv_ld_as_needed): Set
11827 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11828 Use -z ignore, -z record on *-*-solaris2*.
11829 (HAVE_LD_AS_NEEDED): Update comment.
11830 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11831 * configure: Regenerate.
11832 * config.in: Regenerate.
11833 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11834 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11835 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11836 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11837 equivalents. Fix markup.
11838 * doc/tm.texi: Regenerate.
11840 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11842 * config/i386/i386.opt: New option mstack-protector-guard=.
11843 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11844 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11845 TARGET_SSP_TLS_GUARD.
11846 * config/i386/i386.c (ix86_option_override_internal): Set
11847 ix86_stack_protector_guard.
11848 * config/i386/i386.md (stack_protect_set): Enable for
11849 TARGET_SSP_TLS_GUARD only.
11850 (stack_protect_set_<mode>): Ditto.
11851 (stack_protect_test): Ditto.
11852 (stack_protect_test_<mode>): Ditto.
11853 * doc/invoke.texi (i386 Option): Document.
11855 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11858 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11859 (S_MODES): Set H_MODE bit.
11860 (SF_MODES): Set only S_MODE and SF_MODE bits.
11861 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11862 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11863 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11864 <MODE_FLOAT>: Likewise.
11866 2013-04-15 Joey Ye <joey.ye@arm.com>
11868 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11870 2013-04-15 Joey Ye <joey.ye@arm.com>
11872 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11874 (thumb_far_jump_used_p): Count instruction size and set
11877 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11879 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11880 * resource.c (mark_target_live_regs): Reformat conditional block.
11882 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11884 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11885 notes, they are emitted only just before final.
11886 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11888 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11890 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11891 * cfgrtl.c (delete_insn): Call it here instead.
11892 * lra-spills.c (lra_final_code_change): Use delete_insn.
11893 * haifa-sched.c (sched_remove_insn): Likewise.
11894 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11895 returning to the nop pool.
11896 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11897 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11899 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11901 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11902 * doc/tm.texi: Regenerated.
11904 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11906 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11909 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11911 * df-core.c (df_find_def): Compare register numbers.
11912 (df_find_use): Likewise.
11914 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11917 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11918 lra_in_progress for return.
11920 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11922 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11923 define_insn into define_insn_and_split and emit movsicc patterns.
11925 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11927 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11929 2013-04-12 Richard Biener <rguenther@suse.de>
11931 * tree-pass.h (TODO_do_not_ggc_collect): New.
11932 * passes.c (execute_one_ipa_transform_pass): Honor
11933 TODO_do_not_ggc_collect.
11934 (execute_one_pass): Likewise.
11937 2013-04-10 Richard Biener <rguenther@suse.de>
11939 * passes.c (init_optimization_passes): Remove reload pass.
11940 * ira.c (do_reload): Merge into ...
11942 (rest_of_handle_reload): Remove.
11943 (pass_reload): Likewise.
11944 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11945 of reload for vzeroupper pass placement.
11947 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11949 PR tree-optimization/56918
11950 PR tree-optimization/56920
11951 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11952 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11953 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11954 use wide_mul_with_sign method.
11956 2013-04-12 Richard Biener <rguenther@suse.de>
11958 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11959 not be considered a gimple constant.
11961 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11963 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11964 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11967 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11968 Jakub Jelinek <jakub@redhat.com>
11970 * opts.c: Include diagnostic-color.h.
11971 (common_handle_option): Handle OPT_fdiagnostics_color_.
11972 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11973 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11974 (diagnostic-color.o): New.
11975 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11976 (diagnostic_color_rule): New enum.
11977 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11978 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11979 the location string.
11980 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11981 either NULL, or color kind.
11982 * diagnostic-color.c: New file.
11983 * diagnostic-color.h: New file.
11984 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11986 * doc/invoke.texi (-fdiagnostics-color): Document.
11987 * pretty-print.h (pp_show_color): Define.
11988 (struct pretty_print_info): Add show_color field.
11989 * diagnostic.c: Include diagnostic-color.h.
11990 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11991 macros. Colorize error:, warning: etc. strings and also the location
11993 (diagnostic_show_locus): Colorize the caret line.
11994 * pretty-print.c: Include diagnostic-color.h.
11995 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11996 inside of %< %> quotes or quoted through q format modifier.
11998 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12000 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12002 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12004 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12005 code in CC_NZ mode.
12006 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12009 2013-04-11 Marek Polacek <polacek@redhat.com>
12011 PR tree-optimization/48184
12012 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12014 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12016 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12017 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12018 (skip_simple_arithmetic): Tidy up.
12019 * tree.h (skip_simple_constant_arithmetic): Declare.
12021 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12023 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12025 2013-04-11 Richard Biener <rguenther@suse.de>
12027 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12028 generate vector constants.
12030 2013-04-11 Richard Biener <rguenther@suse.de>
12032 PR tree-optimization/56878
12033 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12034 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12036 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12037 Prefer to align the DR with the most invariant base address.
12039 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12041 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12044 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12046 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12047 floating-point vector comparisons against 0.
12049 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12051 PR tree-optimization/56899
12052 * fold-const.c (extract_muldiv_1): Apply distributive law
12053 only if TYPE_OVERFLOW_WRAPS (ctype).
12055 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12058 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12059 loading has side effect.
12061 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12063 * config/sparc/sparc.c: Include tree-pass.h.
12064 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12065 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12066 head of file. Change return type. Split off gate function.
12067 (sparc_gate_work_around_errata): New function.
12068 (pass_work_around_errata): New pass definition.
12069 (insert_pass_work_around_errata) New pass insert definition to
12070 insert pass_work_around_errata just after delayed-branch scheduling.
12071 (sparc_option_override): Insert the pass.
12072 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12074 2013-04-10 David S. Miller <davem@davemloft.net>
12076 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12077 or -mcpu=hypersparc.
12079 * target.def (cstore_mode): New hook.
12080 * target.h: Include insn-codes.h
12081 * targhooks.c: Likewise.
12082 (default_cstore_mode): New function.
12083 * targhooks.h: Declare it.
12084 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12085 * doc/tm.texi: Rebuild.
12086 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12087 target hook, rather than inspecting the insn_data.
12088 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12089 (TARGET_CSTORE_MODE): Redefine.
12090 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12092 * config/sparc/predicates.md (cstore_result_operand): New special
12094 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12095 Use it for operand 0.
12096 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12097 (*snesi_special): Likewise.
12098 (*snesi_zero): Likewise.
12099 (*seqsi_zero): Likewise.
12100 (*sltu_insn): Likewise.
12101 (*sgeu_insn): Likewise.
12102 (*seqdi_special): Make operand 0 and comparison operation be of
12104 (*snedi_special): Likewise.
12105 (*snedi_special_vis3): Likewise.
12106 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12107 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12108 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12109 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12110 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12111 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12112 (*sltu_extend_sp64): Likewise.
12113 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12114 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12115 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12116 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12117 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12119 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12121 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12122 (aarch64_start_file): Use the new function.
12124 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12125 Jason Merrill <jason@redhat.com>
12127 * common.opt: Add -gdwarf.
12128 * opts.c (common_handle_option): Handle it.
12129 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12131 2013-04-10 Richard Biener <rguenther@suse.de>
12133 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12134 (execute_one_ipa_transform_pass): But unconditionally here.
12135 (execute_one_pass): And here.
12136 (init_optimization_passes): Remove reload pass.
12137 * tree-pass.h (TODO_ggc_collect): Remove.
12138 (pass_reload): Likewise.
12139 * ira.c (do_reload): Merge into ...
12141 (rest_of_handle_reload): Remove.
12142 (pass_reload): Likewise.
12143 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12144 of reload for vzeroupper pass placement.
12145 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12146 and todo_flags_finish of all passes.
12148 2013-04-10 Richard Biener <rguenther@suse.de>
12150 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12151 first_const_oprnd field, rename first_def_type to first_op_type.
12152 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12153 (vect_get_and_check_slp_defs): Always use the type of the
12154 operand. Allow mixed vect_external_def, vect_constant_def types.
12155 (vect_get_constant_vectors): Handle mixed vect_external_def,
12156 vect_constant_def types.
12158 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12160 PR tree-optimization/55524
12161 * tree-ssa-math-opts.c
12162 (convert_mult_to_fma): Don't use an fms construct
12163 when we don't have an fms operation, but fnma, and it looks
12164 likely that we'll be able to use the latter.
12166 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12168 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12170 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12171 inline fail caused by overwritable functions.
12173 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12175 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12176 unnecessary bits in the constant power of two case.
12178 2013-04-10 Richard Biener <rguenther@suse.de>
12180 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12181 broken code swapping operands.
12182 (vect_build_slp_tree): Do not compute load permutations here.
12183 (vect_analyze_slp_instance): Compute load permutations here,
12184 after building the SLP tree.
12186 2013-04-09 Christian Bruel <christian.bruel@st.com>
12188 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12189 of next/prev_real_insn.
12191 2013-04-09 Jan Hubicka <jh@suse.cz>
12193 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12194 Drop aliased parameter.
12195 (function_and_variable_visibility): Do not handle alias pairs.
12196 * cgraph.c (varpool_externally_visible_p): Update prototype.
12197 * varpool.c (varpool_add_new_variable): Update.
12199 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12201 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
12203 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
12205 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
12207 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
12209 2013-04-09 Marek Polacek <polacek@redhat.com>
12211 PR tree-optimization/48762
12212 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
12214 2013-04-09 Richard Biener <rguenther@suse.de>
12216 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
12218 (vect_build_slp_tree): Likewise.
12219 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
12220 calculating the cost of a SLP instance.
12221 (vect_analyze_slp_instance): Use it from here, after building
12224 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12226 PR middle-end/56883
12227 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
12228 expand_omp_for_static_chunk): Use simple_p = true in
12229 force_gimple_operand_gsi calls when assigning to addressable decls.
12231 2013-04-09 Jeff Law <law@redhat.com>
12233 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
12234 when the boolean was created by converting a wider object which
12235 had a boolean range.
12237 2013-04-09 Richard Biener <rguenther@suse.de>
12239 * tree-vectorizer.h (slp_void_p): Remove.
12240 (slp_tree): Typedef before _slp_tree declaration.
12241 (struct _slp_tree): Use a vector of slp_tree as children.
12242 (vect_get_place_in_interleaving_chain): Remove.
12243 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
12245 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
12247 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
12248 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
12249 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
12250 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
12251 Use slp_node instead of slp_void_p and adjust.
12253 2013-04-09 Richard Biener <rguenther@suse.de>
12255 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
12256 work that is not necessary.
12258 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12260 PR tree-optimization/56854
12261 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
12262 forward into clobber stmts if it would change MEM_REF lhs into
12265 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
12267 * tree.c (type_hash_lookup, type_hash_add): Make static.
12268 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
12270 2013-04-09 Richard Biener <rguenther@suse.de>
12272 * tree.h (unsave_expr_now): Remove.
12273 * tree-inline.c (mark_local_for_remap_r): Remove.
12274 (unsave_expr_1): Likewise.
12275 (unsave_r): Likewise.
12276 (unsave_expr_now): Likewise.
12277 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
12278 (propagate_tree_value): Likewise.
12280 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
12282 * doc/rtl.texi (sequence): Rewrite documentation to match the
12283 current use of SEQUENCE rtl objects.
12284 * rtl.def (SEQUENCE): Likewise.
12286 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
12287 Update documentation.
12288 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
12289 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
12291 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
12293 2013-04-08 Teresa Johnson <tejohnson@google.com>
12295 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
12296 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
12298 (estimate_edge_size_and_time): Add comment to suggest using rounding
12300 (estimate_node_size_and_time): Ditto.
12301 (remap_edge_change_prob): Use helper rounding divide methods.
12302 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
12303 (gimple_mod_pow2_value_transform): Ditto.
12304 (gimple_mod_subtract_transform): Ditto.
12305 (gimple_ic_transform): Ditto.
12306 (gimple_stringops_transform): Ditto.
12307 * stmt.c (conditional_probability): Ditto.
12308 (emit_case_dispatch_table): Ditto.
12309 * lto-cgraph.c (merge_profile_summaries): Ditto.
12310 * tree-optimize.c (execute_fixup_cfg): Ditto.
12311 * cfgcleanup.c (try_forward_edges): Ditto.
12312 * cfgloopmanip.c (scale_loop_profile): Ditto.
12314 (duplicate_loop_to_header_edge): Ditto.
12315 (lv_adjust_loop_entry_edge): Ditto.
12316 * tree-vect-loop.c (vect_transform_loop): Ditto.
12317 * profile.c (compute_branch_probabilities): Ditto.
12318 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
12319 * lto-streamer-in.c (input_cfg): Ditto.
12320 * gimple-streamer-in.c (input_bb): Ditto.
12321 * ipa-cp.c (update_profiling_info): Ditto.
12322 (update_specialized_profile): Ditto.
12323 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
12324 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
12326 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12327 (compute_trg_info): Ditto.
12328 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
12329 (purge_dead_edges): Ditto.
12330 * loop-unswitch.c (unswitch_loop): Ditto.
12331 * cgraphclones.c (cgraph_clone_edge): Ditto.
12332 (cgraph_clone_node): Ditto.
12333 * tree-inline.c (copy_bb): Ditto.
12334 (copy_edges_for_bb): Ditto.
12335 (initialize_cfun): Ditto.
12336 (copy_cfg_body): Ditto.
12337 (expand_call_inline): Ditto.
12339 2013-04-08 Kai Tietz <ktietz@redhat.com>
12341 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
12342 TARGET_CYGWIN64 by TARGET_64BIT.
12344 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12346 * config/epiphany/epiphany.md (GPR_1): New constant.
12347 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
12348 * config/epiphany/epiphany.c (gen_compare_reg):
12349 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
12350 is already in place.
12351 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
12352 Don't require being called during rtl expansion; If y operlaps r0,
12354 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
12355 (epiphany_expand_epilogue): Likewise.
12357 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
12358 Don't use CC_FPmode for ORDERED / UNORDERED.
12359 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
12361 * config/epiphany/constraints.md (CnL): New constraint.
12362 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
12363 * config/epiphany/predicates.md (add_operand): Allow 1024.
12365 * config/epiphany/epiphany.md (logical_op): New code iterator.
12366 (op_mnc): New code attribute.
12367 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
12368 (mov_f+1, mov_f+2): New peephole2 patterns.
12370 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
12371 (cstoresi4): Also allow re-use of zero result when doing a NE
12372 comparison to a non-zero operand.
12373 Use (clobber (scratch)) for first insn if the gpr output is not needed.
12375 * config/epiphany/epiphany.md (<insn_opname>v2si3):
12376 Use gen_addsi3_i / gen_subsi3_i.
12378 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12382 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
12383 contain anything but clobbers, at most one __builtin_stack_restore,
12384 optionally debug stmts and final resx, and if it has at least one
12385 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
12386 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
12387 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
12388 which isn't defaut definition, remove them.
12389 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
12390 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
12391 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
12392 with MEM_REF LHS with SSA_NAME address.
12394 2013-04-08 Jeff Law <law@redhat.com>
12396 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
12398 2013-04-08 Richard Biener <rguenther@suse.de>
12400 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
12402 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
12403 determined vector type.
12404 (vect_analyze_data_refs): Likewise.
12405 (vect_get_new_vect_var): Adjust.
12406 (vect_create_destination_var): Preserve SSA name versions.
12407 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
12408 not dump anything here.
12410 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12412 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
12413 Add member lr_slot_known.
12414 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
12416 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12417 Remove code that sets lr_slot_offset according to what a previous
12418 version of epiphany_emit_save_restore used to do.
12419 (epiphany_emit_save_restore): When doing an lr save or restore,
12420 set/verify lr_slot_known and lr_slot_offset.
12422 2013-04-08 Xinyu Qi <xyqi@marvell.com>
12425 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
12428 2013-04-08 Richard Biener <rguenther@suse.de>
12430 * alias.c (find_base_term): Fix thinko in previous change.
12432 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12434 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
12435 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
12436 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
12437 if possible to compute val.
12438 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
12439 For QImode integers don't require anything about precision. Use
12440 const_with_all_bytes_same to find out if the constant doesn't have
12441 repeated bytes in it.
12443 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12445 * config/s390/s390.c (s390_expand_insv): Only accept insertions
12448 2013-04-08 Marek Polacek <polacek@redhat.com>
12450 PR rtl-optimization/48182
12451 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
12454 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12457 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
12458 nuses, make sure we have a label.
12460 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12463 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12464 (rs6000_emit_swdiv_low_precision): Remove.
12465 (rs6000_emit_swdiv): Rewrite to handle between one and four
12466 iterations of Newton-Raphson generally; modify required number of
12467 iterations for some cases.
12468 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12470 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12472 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12473 set-but-unused variable.
12475 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12476 basic blocks of released function bodies garbage-collectable.
12478 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12479 (struct rtl_opt_pass): Add TODO_df_finish.
12481 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12483 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12485 * config/arm/constraints.md (q): New constraint.
12486 * config/arm/ldrdstrd.md: New file.
12487 * config/arm/arm.md (ldrdstrd.md) New include.
12488 (arm_movdi): Use "q" instead of "r" constraint
12489 for double-word memory access.
12490 (movdf_soft_insn): Likewise.
12491 * config/arm/vfp.md (movdi_vfp): Likewise.
12492 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12493 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12494 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12495 (mem_ok_for_ldrd_strd): Likewise.
12496 (output_move_double): Update assertion.
12498 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12500 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12502 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12504 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12505 define_insn_and_split.
12506 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12508 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12510 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12511 define_insn_and_split.
12512 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12513 (shiftsi3_compare): New pattern.
12514 (rrx): New pattern.
12515 * config/arm/unspecs.md (UNSPEC_RRX): New.
12517 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12519 * config/arm/arm.md (negdi_extendsidi): New pattern.
12520 (negdi_zero_extendsidi): Likewise.
12522 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12524 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12525 define_insn_and_split.
12526 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12527 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12529 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12531 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12532 define_insn_and_split.
12533 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12534 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12536 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12538 * config/arm/arm.md (subsi3_carryin): New pattern.
12539 (subsi3_carryin_const): Likewise.
12540 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12541 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12543 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12545 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12547 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12549 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12550 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12552 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12554 * config/arm/arm.c (arm_expand_builtin): Change fcode
12555 type to unsigned int.
12557 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12559 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12561 2013-04-04 Ian Lance Taylor <iant@google.com>
12563 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12566 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12568 PR middle-end/56729
12569 * df-scan.c (df_insn_delete): Disable failing assert.
12571 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12573 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12574 New function prototype.
12575 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12576 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12577 (arm_builtin_vectorized_function): New function.
12579 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12581 * config/arm/arm_neon_builtins.def: New file.
12582 * config/arm/arm.c (neon_builtin_data): Move contents to
12583 arm_neon_builtins.def.
12584 (enum arm_builtins): Include neon builtin definitions.
12585 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12586 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12588 2013-04-04 Marek Polacek <polacek@redhat.com>
12590 PR tree-optimization/48186
12591 * predict.c (maybe_hot_frequency_p): Return false if
12592 HOT_BB_FREQUENCY_FRACTION is 0.
12593 (cgraph_maybe_hot_edge_p): Likewise.
12595 2013-04-04 Richard Biener <rguenther@suse.de>
12597 PR tree-optimization/56826
12598 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12601 2013-04-04 Richard Biener <rguenther@suse.de>
12603 PR tree-optimization/56213
12604 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12605 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12607 2013-04-04 Richard Biener <rguenther@suse.de>
12609 PR tree-optimization/56837
12610 * tree-loop-distribution.c (classify_partition): For non-zero
12611 values require that the value has the same precision as its
12612 mode to be useful as memset value.
12614 2013-04-03 Nick Clifton <nickc@redhat.com>
12616 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12617 (fmssf4): Use fmsf.s on E3V5 architectures.
12618 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12619 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12621 2013-04-03 Jeff Law <law@redhat.com>
12623 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12624 (lra-eliminations.o): Likewise.
12626 2013-04-03 Teresa Johnson <tejohnson@google.com>
12628 * gcov-io.c (compute_working_sets): Moved most of body of old
12629 compute_working_sets here from profile.c.
12630 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12631 (gcov_working_set_t): Moved typedef here from basic-block.h
12632 (compute_working_set): Declare.
12633 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12634 (get_working_sets): Renamed from compute_working_set,
12635 replace most of body with call to new compute_working_sets.
12636 (get_exec_counts): Replace call to compute_working_sets
12637 to get_working_sets.
12638 * profile.h (get_working_sets): Renamed from compute_working_set.
12639 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12640 to get_working_sets.
12641 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12642 * gcov-dump.c (dump_working_sets): New function.
12644 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12646 * hwint.c (sext_hwi, zext_hwi): New functions.
12647 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12648 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12649 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12650 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12651 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12652 (sext_hwi, zext_hwi): New functions.
12654 2013-04-03 Jeff Law <law@redhat.com>
12656 PR tree-optimization/56799
12657 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12658 back test for widening conversion erroneously dropped in prior change.
12660 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12663 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12664 instead of next_real_insn.
12666 2013-04-03 Marek Polacek <polacek@redhat.com>
12669 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12671 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12674 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12676 (thumb1_output_casesi): Likewise.
12677 (thumb2_output_casesi): Likewise.
12679 2013-04-03 Richard Biener <rguenther@suse.de>
12681 PR tree-optimization/56817
12682 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12684 (tree_unroll_loops_completely_1): ... new function to manually
12685 walk the loop tree, properly defering outer loops of unrolled
12686 loops to later iterations.
12688 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12690 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12691 (vectorizable_load): Likewise.
12692 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12693 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12695 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12697 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12700 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12702 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12704 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12706 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12708 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12710 PR tree-optimization/56790
12711 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12714 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12716 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12718 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12719 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12722 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12725 * tree.h (force_folding_builtin_constant_p): New decl.
12726 * builtins.c (force_folding_builtin_constant_p): New variable.
12727 (fold_builtin_constant_p): Fold immediately also if
12728 force_folding_builtin_constant_p.
12730 2013-04-03 Richard Biener <rguenther@suse.de>
12732 PR tree-optimization/56812
12733 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12734 DRs of the same interleaving chain are independent.
12736 2013-04-02 Jason Merrill <jason@redhat.com>
12738 * gdbinit.in (pbb): Use debug fn.
12740 2013-04-02 Lawrence Crowl <crowl@google.com>
12742 * sese.h (struct ivtype_map_elt_s): Remove unused.
12743 (extern debug_ivtype_map): Remove unused.
12744 (extern eq_ivtype_map_elts): Remove unused.
12745 * sese.c (debug_ivtype_map): Removed unused.
12746 (debug_ivtype_map_1): Removed unused.
12747 (debug_ivtype_elt): Remove unused.
12748 (eq_ivtype_map_elts): Remove unused.
12751 2013-04-02 Kai Tietz <ktietz@redhat.com>
12754 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12755 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12756 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12758 (legitimize_pe_coff_symbol): Likewise.
12759 (is_imported_p): New helper-function.
12760 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12761 for Windows x64 targets.
12762 (ix86_expand_prologue): Optimize for pe-coff targets.
12763 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12764 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12765 medium/large code-model.
12766 (legitimize_pic_address): Likewise.
12767 (legitimize_tls_address): Likewise.
12768 (ix86_expand_call): Likewise.
12769 (x86_output_mi_thunk): Likewise.
12770 (get_dllimport_decl): Add new beimport argument.
12771 (construct_plt_address): Don't assert for x64 pe-coff targets.
12772 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12774 (SYMBOL_FLAG_STUBVAR): New macro.
12775 (SYMBOL_REF_STUBVAR_P): Likewise.
12776 * config/i386/winnt.c (stub_list): New structure.
12777 (stub_head): New local variable.
12778 (i386_pe_record_stub): New function.
12779 (i386_pe_file_end): Emit refptr-stubs.
12781 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12783 PR rtl-optimization/56745
12784 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12785 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12788 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12789 and both of them are MEM_REFs, just compare first argument for
12790 equality and attempt to deal even with differing offsets.
12793 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12794 of gimple_clobber_p to be MEM_REF.
12795 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12796 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12797 after gimplification.
12798 * asan.c (get_mem_ref_of_assignment): Don't instrument
12799 gimple_clobber_p stmts.
12800 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12801 gimple_clobber_p stmt if they have MEM_REF lhs and
12802 are dead because of another gimple_clobber_p stmt.
12803 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12804 gimple_clobber_p stmts like debug stmts.
12805 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12806 that refer to unused VAR_DECLs or uninitialized values.
12807 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12808 gimple_clobber_p stmts if they refer to removed parameters.
12809 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12812 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12814 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12815 using SWI48 mode attribute.
12817 2013-04-02 Wei Mi <wmi@google.com>
12819 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12820 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12821 *<rotate_insn><mode>3_mask in i386.md.
12823 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12825 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12827 2013-04-02 Richard Biener <rguenther@suse.de>
12829 PR tree-optimization/56778
12830 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12831 Runtime alias tests are not supported for gather loads.
12832 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12833 stmts referenced from SSA operands before updating SSA form.
12835 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12836 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12838 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12839 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12840 * config/arm/cortex-a53.md: New file.
12841 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12842 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12843 * config/arm/arm.c (arm_issue_rate): Likewise.
12844 * config/arm/arm-tune.md: Regenerate
12845 * config/arm/arm-tables.opt: Regenerate.
12846 * config/arm/arm-cores.def: Add cortex-a53.
12848 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12850 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12853 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12855 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12856 scalar load/store operations using B/H registers.
12857 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12859 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12861 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12863 * config/aarch64/aarch64.c
12864 (aarch64_simd_scalar_immediate_valid_for_move): New.
12865 * config/aarch64/aarch64-protos.h
12866 (aarch64_simd_scalar_immediate_valid_for_move): New.
12867 * config/aarch64/constraints.md (Dh, Dq): New.
12868 * config/aarch64/iterators.md (hq): New.
12870 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12872 * reorg.c (get_branch_condition): Deal with conditional returns.
12873 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12875 2013-04-01 Wei Mi <wmi@google.com>
12877 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12878 Truncate operand 2 using %b asm operand modifier.
12879 (*<shift_insn><mode>3_mask): Ditto.
12880 (*<rotate_insn><mode>3_mask): Ditto.
12882 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12884 PR middle-end/56798
12885 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12887 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12889 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12891 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12893 2013-03-30 Lawrence Crowl <crowl@google.com>
12895 * dse.c (clear_alias_sets): Remove never set.
12896 (disqualified_clear_alias_sets): Remove never set.
12897 (clear_alias_mode_pool): Remove never set.
12898 (dse_step0): Remove condition that is never true.
12899 (canon_address): Remove condition that is never true.
12900 (dse_step7): Remove condition that is never true.
12901 (rest_of_handle_dse): Remove condition that is never true.
12902 (rest_of_handle_dse::did_global): Remove never read from above.
12903 (dse_step2_spill): Remove never called from above.
12904 (dse_step5_spill): Remove never called from above.
12906 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12908 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12909 JUMP_TABLE_DATA changes.
12910 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12911 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12912 (Insns) <jump_table_data>: New entry.
12913 * doc/tm.texi: Regenerate.
12915 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12917 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12918 for table jump at the end of a basic block using tablejump_p.
12919 * targhooks.c (default_invalid_within_doloop): Likewise.
12920 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12921 target hook implementation that is identical to the default hook.
12922 (rs6000_invalid_within_doloop): Remove.
12924 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12925 unused variable from tablejump_p call.
12927 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12928 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12929 (INSN_DELETED_P): Likewise.
12930 (emit_jump_table_data): New prototype.
12931 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12932 after 4th as unused.
12933 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12934 * sched-vis.c (print_insn): Likewise.
12935 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12936 insn for compatibility with back ends that use next_active_insn to
12937 identify jump table data.
12938 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12939 (remove_insn): Likewise.
12940 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12942 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12943 (emit_jump_table_data): New function.
12945 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12946 basic block, a JUMP_TABLE_DATA never is.
12947 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12948 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12949 off from code handling real insns.
12950 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12951 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12952 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12953 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12954 is not a NONDEBUG_INSN_P.
12955 * ira-costs.c (scan_one_insn): Likewise.
12956 * jump.c (mark_all_labels): Likewise.
12957 (mark_jump_label_1): Likewise.
12958 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12959 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12961 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12962 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12963 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12964 JUMP_TABLE_DATA_P insns.
12965 (calculate_elim_costs_all_insns): Likewise.
12966 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12967 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12968 (delete_output_reload): Code style fixups.
12969 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12970 insn flags on this non-insn.
12971 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12972 as scheduling barriers, for pre-change compatibility.
12973 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12974 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12976 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12977 redundant JUMP_TABLE_DATA_P test.
12978 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12979 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12980 (frv_for_each_packet): Likewise.
12981 * config/i386/i386.c (min_insn_size): Likewise.
12982 (ix86_avoid_jump_mispredicts): Likewise.
12983 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12984 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12985 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12986 (mips16_insn_length): Robustify.
12987 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12988 (mips16_split_long_branches): Likewise.
12989 * config/pa/pa.c (pa_combine_instructions): Likewise.
12990 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12991 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12992 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12993 as contributing to pool range lengths.
12994 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12995 Remove redundant JUMP_TABLE_DATA_P test.
12996 (sh_loop_align): Likewise.
12997 (split_branches): Likewise.
12998 (sh_insn_length_adjustment): Likewise.
12999 * config/spu/spu.c (get_branch_target): Likewise.
13001 2013-03-29 Jan Hubicka <jh@suse.cz>
13003 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13004 gcov streaming; stream hot bb threshold to ltrans.
13005 * predict.c (get_hot_bb_threshold): Break out from ....
13006 (maybe_hot_count_p): ... here.
13007 (set_hot_bb_threshold): New function.
13008 * lto-section-in.c (lto_section_name): Add profile.
13009 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13010 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13011 and data-streamer.h
13012 (histogram_entry): New structure.
13013 (histogram, histogram_pool): New global vars.
13014 (histogram_hash): New structure.
13015 (histogram_hash::hash): New method.
13016 (histogram_hash::equal): Likewise.
13017 (account_time_size): New function.
13018 (cmp_counts): New function.
13019 (dump_histogram): New function.
13020 (ipa_profile_generate_summary): New function.
13021 (ipa_profile_write_summary): New function.
13022 (ipa_profile_read_summary): New function.
13023 (ipa_profile): Decide on threshold.
13024 (pass_ipa_profile): Add ipa_profile_write_summary and
13025 ipa_profile_read_summary.
13026 * Makefile.in (ipa.o): Update dependencies.
13027 * lto-streamer.h (LTO_section_ipa_profile): New section.
13029 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13031 * tree.h (VAR_P): New.
13033 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13036 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13038 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13040 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13041 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13042 (control_flow_insn_p): Likewise.
13043 * cfgrtl.c (duplicate_insn_chain): Likewise.
13044 * final.c (get_attr_length_1): Likewise.
13045 (shorten_branches): Likewise.
13046 (final_scan_insn): Likewise.
13047 * function.c (instantiate_virtual_regs): Likewise.
13048 * gcse.c (insert_insn_end_basic_block): Likewise.
13049 * ira-costs.c (scan_one_insn): Likewise.
13050 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13051 * lra.c (check_rtl): Likewise.
13052 * reload1.c (elimination_costs_in_insn): Likewise.
13053 * reorg.c (follow_jumps): Likewise.
13055 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13056 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13057 (thumb_far_jump_used_p): Likewise.
13058 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13059 (workaround_speculation): Likewise.
13060 (add_sched_insns_for_speculation): Likewise.
13061 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13062 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13063 (frv_for_each_packet): Likewise.
13064 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13065 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13066 (final_emit_insn_group_barriers): Likewise.
13067 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13068 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13069 (mips16_insn_length): Likewise.
13070 * config/pa/pa.c (pa_reorg): Likewise.
13071 (pa_combine_instructions): Likewise.
13072 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13073 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13074 (sh_reorg): Likewise.
13075 (split_branches): Likewise.
13076 * config/spu/spu.c (get_branch_target): Likewise.
13078 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13081 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13083 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13084 Fix declaration name.
13086 2013-03-28 Lawrence Crowl <crowl@google.com>
13088 * graphds.h (struct graph.indicies): Remove unused.
13089 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13090 (SCOP_ORIGINAL_PDDRS): Remove unused.
13091 * sese.h (extern insert_loop_close_phis): Removed unused.
13092 (extern insert_guard_phis): Removed unused.
13093 (extern ivtype_map_elt_info): Removed unused.
13094 (new_ivtype_map_elt): Removed unused.
13095 * sese.c (ivtype_map_elt_info): Removed unused.
13097 2013-03-28 Lawrence Crowl <crowl@google.com>
13099 * Makefile.in: Add several missing include dependences.
13101 (test-dump.o): New. This object is not added to any executable,
13102 but is present for ad-hoc testing.
13104 (debug (const bitmap_head_def &)): New.
13105 (debug (const bitmap_head_def *)): New.
13107 (extern debug (const bitmap_head_def &)): New.
13108 (extern debug (const bitmap_head_def *)): New.
13110 (debug (edge_def &)): New.
13111 (debug (edge_def *)): New.
13113 (debug (basic_block_def &)): New.
13114 (debug (basic_block_def *)): New.
13116 (dump_node (const_tree, int, FILE *)): Correct source file.
13118 (debug (die_struct &)): New.
13119 (debug (die_struct *)): New.
13121 (extern debug (die_struct &)): New.
13122 (extern debug (die_struct *)): New.
13123 * gimple-pretty-print.c
13124 (debug (gimple_statement_d &)): New.
13125 (debug (gimple_statement_d *)): New.
13126 * gimple-pretty-print.h
13127 (extern debug (gimple_statement_d &)): New.
13128 (extern debug (gimple_statement_d *)): New.
13130 (debug (ira_allocno_copy &)): New.
13131 (debug (ira_allocno_copy *)): New.
13132 (debug (ira_allocno &)): New.
13133 (debug (ira_allocno *)): New.
13135 (extern debug (ira_allocno_copy &)): New.
13136 (extern debug (ira_allocno_copy *)): New.
13137 (extern debug (ira_allocno &)): New.
13138 (extern debug (ira_allocno *)): New.
13140 (debug (live_range &)): New.
13141 (debug (live_range *)): New.
13143 (debug (lra_live_range &)): New.
13144 (debug (lra_live_range *)): New.
13146 (debug (lra_live_range &)): New.
13147 (debug (lra_live_range *)): New.
13149 (debug (omega_pb_d &)): New.
13150 (debug (omega_pb_d *)): New.
13152 (extern debug (omega_pb_d &)): New.
13153 (extern debug (omega_pb_d *)): New.
13155 (debug (const rtx_def &)): New.
13156 (debug (const rtx_def *)): New.
13158 (debug_tree (tree): Move within file.
13159 (debug_raw (const tree_node &)): New.
13160 (debug_raw (const tree_node *)): New.
13161 (dump_tree_via_hooks (const tree_node *, int)): New.
13162 (debug (const tree_node &)): New.
13163 (debug (const tree_node *)): New.
13164 (debug_verbose (const tree_node &)): New.
13165 (debug_verbose (const tree_node *)): New.
13166 (debug_head (const tree_node &)): New.
13167 (debug_head (const tree_node *)): New.
13168 (debug_body (const tree_node &)): New.
13169 (debug_body (const tree_node *)): New.
13170 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13171 (debug (vec<tree, va_gc> &)): New.
13172 (debug (vec<tree, va_gc> *)): New.
13174 (extern debug (const rtx_def &)): New.
13175 (extern debug (const rtx_def *)): New.
13177 (debug_raw (simple_bitmap_def &)): New.
13178 (debug_raw (simple_bitmap_def *)): New.
13179 (debug (simple_bitmap_def &)): New.
13180 (debug (simple_bitmap_def *)): New.
13182 (extern debug (simple_bitmap_def &)): New.
13183 (extern debug (simple_bitmap_def *)): New.
13184 (extern debug_raw (simple_bitmap_def &)): New.
13185 (extern debug_raw (simple_bitmap_def *)): New.
13187 (debug (vinsn_def &)): New.
13188 (debug (vinsn_def *)): New.
13189 (debug_verbose (vinsn_def &)): New.
13190 (debug_verbose (vinsn_def *)): New.
13191 (debug (expr_def &)): New.
13192 (debug (expr_def *)): New.
13193 (debug_verbose (expr_def &)): New.
13194 (debug_verbose (expr_def *)): New.
13195 (debug (vec<rtx> &)): New.
13196 (debug (vec<rtx> *)): New.
13198 (extern debug (vinsn_def &)): New.
13199 (extern debug (vinsn_def *)): New.
13200 (extern debug_verbose (vinsn_def &)): New.
13201 (extern debug_verbose (vinsn_def *)): New.
13202 (extern debug (expr_def &)): New.
13203 (extern debug (expr_def *)): New.
13204 (extern debug_verbose (expr_def &)): New.
13205 (extern debug_verbose (expr_def *)): New.
13206 (extern debug (vec<rtx> &)): New.
13207 (extern debug (vec<rtx> *)): New.
13209 (_list_iter_cond_expr): Make inline instead of static.
13211 (debug (sreal &)): New.
13212 (debug (sreal *)): New.
13214 (extern debug (sreal &)): New.
13215 (extern debug (sreal *)): New.
13217 (extern debug_raw (const tree_node &)): New.
13218 (extern debug_raw (const tree_node *)): New.
13219 (extern debug (const tree_node &)): New.
13220 (extern debug (const tree_node *)): New.
13221 (extern debug_verbose (const tree_node &)): New.
13222 (extern debug_verbose (const tree_node *)): New.
13223 (extern debug_head (const tree_node &)): New.
13224 (extern debug_head (const tree_node *)): New.
13225 (extern debug_body (const tree_node &)): New.
13226 (extern debug_body (const tree_node *)): New.
13227 (extern debug (vec<tree, va_gc> &)): New.
13228 (extern debug (vec<tree, va_gc> *)): New.
13230 (debug (struct loop &)): New.
13231 (debug (struct loop *)): New.
13232 (debug_verbose (struct loop &)): New.
13233 (debug_verbose (struct loop *)): New.
13234 * tree-dump.c: Add header dependence.
13236 (extern debug (struct loop &)): New.
13237 (extern debug (struct loop *)): New.
13238 (extern debug_verbose (struct loop &)): New.
13239 (extern debug_verbose (struct loop *)): New.
13241 (debug (data_reference &)): New.
13242 (debug (data_reference *)): New.
13243 (debug (vec<data_reference_p> &)): New.
13244 (debug (vec<data_reference_p> *)): New.
13245 (debug (vec<ddr_p> &)): New.
13246 (debug (vec<ddr_p> *)): New.
13248 (extern debug (data_reference &)): New.
13249 (extern debug (data_reference *)): New.
13250 (extern debug (vec<data_reference_p> &)): New.
13251 (extern debug (vec<data_reference_p> *)): New.
13252 (extern debug (vec<ddr_p> &)): New.
13253 (extern debug (vec<ddr_p> *)): New.
13255 (debug (pt_solution &)): New.
13256 (debug (pt_solution *)): New.
13258 (extern debug (pt_solution &)): New.
13259 (extern debug (pt_solution *)): New.
13261 (debug (_var_map &)): New.
13262 (debug (_var_map *)): New.
13263 (debug (tree_live_info_d &)): New.
13264 (debug (tree_live_info_d *)): New.
13266 (extern debug (_var_map &)): New.
13267 (extern debug (_var_map *)): New.
13268 (extern debug (tree_live_info_d &)): New.
13269 (extern debug (tree_live_info_d *)): New.
13271 2013-03-28 Jan Hubicka <jh@suse.cz>
13273 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
13275 2013-03-28 Ian Bolton <ian.bolton@arm.com>
13277 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
13278 record only when desired or required.
13280 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
13282 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
13283 *vec_extractv2di_1_rex64. Use x64 isa attribute.
13285 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13287 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
13288 (*andsi3_compare0_uxtw): New pattern.
13289 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
13290 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
13292 2013-03-28 Jan Hubicka <jh@suse.cz>
13294 * data-streamer-in.c (streamer_read_gcov_count): New function.
13295 * gimple-streamer-out.c: Include value-prof.h.
13296 (output_gimple_stmt): Output histogram.
13297 (output_bb): Use streamer_write_gcov_count.
13298 * value-prof.c: Include data-streamer.h
13299 (dump_histogram_value): Add HIST_TYPE_MAX.
13300 (stream_out_histogram_value): New function.
13301 (stream_in_histogram_value): New function.
13302 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
13303 (stream_out_histogram_value, stream_in_histogram_value): Declare.
13304 * data-streamer-out.c (streamer_write_gcov_count): New function.
13305 (streamer_write_gcov_count_stream): New function.
13306 * lto-cgraph.c (lto_output_edge): Update counter streaming.
13307 (lto_output_node): Likewise.
13308 (input_node, input_edge): Likewise.
13309 * lto-streamer-out.c (output_cfg): Update streaming.
13310 * lto-streamer-in.c (input_cfg): Likewise.
13311 * data-streamer.h (streamer_write_gcov_count,
13312 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
13313 * gimple-streamer-in.c: Include value-prof.h
13314 (input_gimple_stmt): Input histograms.
13315 (input_bb): Update profile streaming.
13317 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13319 * genmodes.c (emit_max_int): New function.
13320 (emit_insn_modes_h): Added call to emit_max_function.
13321 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
13323 * machmode.def: Fixed comment.
13325 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13327 * combine.c (try_combine): Removed useless assert.
13328 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
13330 2013-03-28 Marek Polacek <polacek@redhat.com>
13331 Richard Biener <rguenther@suse.de>
13333 PR tree-optimization/56695
13334 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
13335 build signed result of a vector comparison.
13336 * tree-cfg.c (verify_gimple_comparison): Check that a result
13337 of a vector comparison has signed type.
13339 2013-03-28 Richard Biener <rguenther@suse.de>
13341 PR tree-optimization/37021
13342 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
13343 do not restrict gaps between groups.
13344 * tree-vect-stmts.c (vectorizable_load): Properly account for
13345 a gap between groups.
13347 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
13349 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
13350 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
13353 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
13355 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
13356 * doc/extend.texi (Named Address Spaces): Ditto.
13357 (Variable Attributes): Ditto.
13359 2013-03-27 Kai Tietz <ktietz@redhat.com>
13361 * config.build: Add support for cygwin x64 target.
13362 * config.gcc: Likewise.
13363 * config.host: Likewise.
13364 * configure.ac: Likewise
13365 * configure: Regenerated.
13367 2013-03-27 Kai Tietz <ktietz@redhat.com>
13369 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
13370 * config/i386/t-cygwin-w64: New file.
13371 * config/i386/cygwin-w64.h: New file.
13372 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
13373 and add support for x64-cygwin target.
13374 (CPP_SPEC): Likewise.
13375 (CXX_WRAP_SPEC_LIST): Undefine before define.
13376 (LIBGCJ_SONAME): Use 15 as version.
13378 2013-03-27 Richard Biener <rguenther@suse.de>
13380 PR tree-optimization/56716
13381 * tree-ssa-structalias.c (perform_var_substitution): Adjust
13382 dumping for ref nodes.
13384 2013-03-27 Martin Jambor <mjambor@suse.cz>
13386 PR tree-optimization/55334
13387 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
13388 restricted pointers to arrays.
13390 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
13392 * Makefile.in (.SUFFIXES): Add .cc.
13393 (.c.o): Apply same recipe for implicit rule .cc.o.
13395 2013-03-27 Richard Biener <rguenther@suse.de>
13397 PR tree-optimization/37021
13398 * tree-vect-data-refs.c (vect_check_strided_load): Allow
13399 REALPART/IMAGPART_EXPRs around the supported refs.
13400 * tree-ssa-structalias.c (find_func_aliases): Assume that
13401 floating-point values are not used to transfer pointers.
13403 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
13405 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
13406 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
13407 * gcc/doc/tm.texi: Regenerate.
13408 * gcc/targhooks.h (default_has_ifunc_p): New.
13409 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
13410 * gcc/config/linux-protos.h: New file.
13411 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
13412 this hook for linux which disables support of indirect functions in
13414 * gcc/config/linux-android.c: New file.
13415 * gcc/config/t-linux-android.c: Ditto.
13416 * gcc/config.gcc: Added new object file linux-android.o.
13417 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
13418 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
13419 * gcc/varasm.c (do_assemble_alias): Likewise.
13420 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
13421 doesn't support indirect functions.
13422 * configure: Regenerate.
13424 2013-03-27 Bin Cheng <bin.cheng@arm.com>
13427 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
13428 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
13431 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13433 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
13435 2013-03-27 Terry Guo <terry.guo@arm.com>
13437 * config/arm/arm-cores.def: Added core cortex-r7.
13438 * config/arm/arm-tune.md: Regenerated.
13439 * config/arm/arm-tables.opt: Regenerated.
13440 * doc/invoke.texi: Added entry for core cortex-r7.
13442 2013-03-27 Walter Lee <walt@tilera.com>
13444 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
13445 double-decrement of next_scratch_regno.
13447 2013-03-27 Walter Lee <walt@tilera.com>
13449 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
13451 (insn_v1mulus): Ditto.
13452 (insn_v2muls): Ditto.
13454 2013-03-27 Walter Lee <walt@tilera.com>
13456 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
13457 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
13459 2013-03-27 Walter Lee <walt@tilera.com>
13461 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13462 (*sibcall_value): Ditto.
13464 2013-03-27 Walter Lee <walt@tilera.com>
13466 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13467 (insn_mnz_v8qi): ... this ...
13468 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13470 (insn_v<n>mnz): Replaced by ...
13471 (insn_v1mnz): ... this ...
13472 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13474 (insn_mz_<mode>): Replaced by ...
13475 (insn_mz_v8qi): ... this ...
13476 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13478 (insn_v<n>mz): Replaced by ...
13479 (insn_v1mz): ... this ...
13480 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13483 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13485 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13487 2013-03-26 Roland McGrath <mcgrathr@google.com>
13489 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13490 than fprintf with a non-constant, non-format string.
13492 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13494 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13495 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13496 operand 0 predicate.
13497 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13498 attribute. Use general_x64nomem_operand as operand 1 predicate.
13499 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13500 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13501 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13502 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13503 (mov<mode>_insv_1): Remove expander. Merge insn with
13504 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13505 Use general_x64nomem_operand as operand 1 predicate.
13506 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13507 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13508 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13509 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13510 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13511 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13512 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13513 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13514 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13515 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13516 (general_x64nomem_operand): Ditto.
13518 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13520 * config/rtems.opt: Add -pthread option.
13522 2013-03-26 Richard Biener <rguenther@suse.de>
13524 * alias.c (find_base_term): Avoid redundant and not used recursion.
13525 (base_alias_check): Get the initial base term from the caller.
13526 (true_dependence_1): Compute and pass base terms to base_alias_check.
13527 (write_dependence_p): Likewise.
13528 (may_alias_p): Likewise.
13530 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13532 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13533 PC-relative load in SI modes and above only.
13535 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13537 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13538 * config/arm/iwmmxt.md (WCGR0): Update.
13539 (WCGR1, WCGR2, WCGR3): Likewise.
13541 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13543 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13544 Use x64 and nox64 isa attributes.
13546 2013-03-26 Richard Biener <rguenther@suse.de>
13548 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13549 alignment computations and rely on get_object_alignment_1
13550 for the !TYPE_P case.
13551 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13553 2013-03-26 Walter Lee <walt@tilera.com>
13555 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13556 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13558 2013-03-25 Jeff Law <law@redhat.com>
13560 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13561 check for INTEGRAL_TYPE_P that was missing due to checking in
13562 wrong version of prior patch.
13564 2013-03-25 Walter Lee <walt@tilera.com>
13566 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13567 TILEGX_INSN_SHUFFLEBYTES1.
13568 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13570 (tilegx_builtins): Ditto.
13571 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13573 2013-03-25 Walter Lee <walt@tilera.com>
13575 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13576 (floatunssisf2): New pattern.
13577 (floatsidf2): New pattern.
13578 (floatunssidf2): New pattern.
13580 2013-03-25 Walter Lee <walt@tilera.com>
13582 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13583 tests for constraint J, K, N, P.
13585 2013-03-25 Walter Lee <walt@tilera.com>
13587 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13588 Use indirect/pcrel encoding.
13589 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13592 2013-03-25 Steve Ellcey <sellcey@mips.com>
13594 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13595 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13596 * config/mips/mips.c (mips_option_override): Set IMADD default.
13597 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13598 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13599 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13600 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13601 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13603 2013-03-25 Jeff Law <law@redhat.com>
13605 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13606 slightly to avoid creating and folding useless trees. Simplify
13607 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13609 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13611 * config/i386/i386.md (*zero_extendsidi2): Merge with
13612 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13613 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13614 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13615 targets. Clarify comment.
13617 2013-03-25 Martin Jambor <mjambor@suse.cz>
13619 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13620 pass-through jump functions differently.
13621 (ipa_read_jump_function): Likewise. Also use setter functions to set
13624 2013-03-25 Martin Jambor <mjambor@suse.cz>
13626 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13627 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13629 (ipa_get_indirect_edge_target): New function.
13630 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13631 ipa_get_indirect_edge_target. Update all callers.
13632 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13633 ipa_get_indirect_edge_target_1 instead of calling
13634 ipa_get_indirect_edge_target.
13635 (create_specialized_node): Pass aggvlas to
13636 ipcp_discover_new_direct_edges.
13638 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13640 * config/arm/arm.md (f_sels, f_seld): New types.
13641 (*cmov<mode>): New pattern.
13642 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13645 2013-03-25 Kai Tietz <ktietz@redhat.com>
13647 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13648 POSIX-printf for mingw-hosted builds.
13650 2013-03-25 Richard Biener <rguenther@suse.de>
13652 PR middle-end/56694
13653 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13654 must-not-throw stmt location.
13656 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13658 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13659 Emit load-acquire versions when acq is true.
13660 (arm_emit_store_exclusive): Add rel parameter.
13661 Emit store-release versions when rel is true.
13662 (arm_split_compare_and_swap): Use acquire-release instructions
13664 of barriers when appropriate.
13665 (arm_split_atomic_op): Likewise.
13666 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13667 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13668 (VUNSPEC_SLX): Likewise.
13669 (VUNSPEC_LDA): Likewise.
13670 (VUNSPEC_STL): Likewise.
13671 * config/arm/sync.md (atomic_load<mode>): New pattern.
13672 (atomic_store<mode>): Likewise.
13673 (arm_load_acquire_exclusive<mode>): Likewise.
13674 (arm_load_acquire_exclusivesi): Likewise.
13675 (arm_load_acquire_exclusivedi): Likewise.
13676 (arm_store_release_exclusive<mode>): Likewise.
13678 2013-03-25 Catherine Moore <clm@codesourcery.com>
13680 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13681 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13682 * config/mip/predicates.md (lwsp_swsp_operand,
13683 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13684 sb16_operand, db4_operand, db7_operand, ib3_operand,
13685 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13686 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13687 andi16_operand): New predicates.
13688 * config/mips/mips.md (compression): New attribute.
13689 (enabled): New attribute.
13690 (length): Consider compression in computing length.
13691 (shift_compression): New code attribute.
13692 (*add<mode>3): New operands. Record compression.
13693 (sub<mode>3): Likewise.
13694 (one_cmpl<mode>2): Likewise.
13695 (*and<mode>3): Likewise.
13696 (*ior<mode>3): Likewise.
13697 (unnamed pattern for xor): Likewise.
13698 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13699 (*<optab><mode>3): Likewise.
13700 (*mov<mode>_internal: Likewise.
13701 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13702 (mips_unsigned_immediate_p): New.
13703 (umips_lwsp_swsp_address_p): New.
13704 (m16_based_address_p): New.
13705 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13706 (mips_unsigned_immediate_p): New prototype.
13707 (lwsp_swsp_address_p): New prototype.
13708 (m16_based_address_p): New prototype.
13709 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13710 (mips_signed_immediate_p): New function.
13711 (m16_based_address_p): New function.
13712 (lwsp_swsp_address_p): New function.
13713 (mips_print_operand_punctuation): Recognize short delay slot insns
13714 for microMIPS.add<mode>3"
13716 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13719 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13720 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13722 2013-03-25 Richard Biener <rguenther@suse.de>
13724 PR tree-optimization/56689
13725 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13728 2013-03-25 Richard Biener <rguenther@suse.de>
13730 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13732 (memory_references): Likewise.
13733 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13734 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13735 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13736 (gather_mem_refs_in_loops): Fold into ...
13737 (analyze_memory_references): ... this. Move initialization
13738 to tree_ssa_lim_initialize.
13739 (fill_always_executed_in): Rename to ...
13740 (fill_always_executed_in_1): ... this.
13741 (fill_always_executed_in): Move contains_call computation to
13742 this new function from ...
13743 (tree_ssa_lim_initialize): ... here.
13744 (tree_ssa_lim): Call fill_always_executed_in.
13746 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13748 * postreload.c (reload_combine): Fix code detecting returns.
13750 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13752 * function.c (emit_use_return_register_into_block): On cc0 targets,
13753 do not emit the sequence between cc0 setter and user.
13755 2013-03-25 Kai Tietz <ktietz@redhat.com>
13757 * config/i386/predicates.md (local_symbolic_operand): Interpret
13758 dll-imported symbols as none-local.
13760 2013-03-25 Richard Biener <rguenther@suse.de>
13762 * tree-ssa-loop-im.c (struct depend): Remove.
13763 (struct lim_aux_data): Make depends a vec of gimples.
13764 (free_lim_aux_data): Adjust.
13765 (add_dependency): Likewise.
13766 (set_level): Likewise.
13768 2013-03-25 Richard Biener <rguenther@suse.de>
13770 PR middle-end/56434
13771 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13772 the pointer returned by calls with ECF_MALLOC set.
13774 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13776 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13778 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13780 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13781 using MMXMODE mode iterator.
13782 (*move<mode>_internal): Merge with *movv2sf_internal and
13783 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13785 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13787 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13788 (record_last_mem_set_info): Likewise.
13790 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13791 of XNEWVEC followed by memset.
13792 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13794 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13796 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13797 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13798 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13799 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13800 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13801 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13802 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13803 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13804 BARRIER_P instead of GET_CODE.
13806 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13808 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13809 inaccuracy in the probing code.
13811 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13812 (ctrapdi4): Likewise.
13814 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13816 * calls.c (expand_call): Add missing guard to code handling return
13817 of non-BLKmode structures in MSB.
13818 * function.c (expand_function_end): Likewise.
13820 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13822 * combine.c (try_combine): Adjust comment. Do not add the set of
13823 insn #0 if the destination indirectly is set or dies in insn #2.
13824 Tidy up code to distribute a new note.
13826 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13828 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13829 also for alternatives 16 and 17.
13831 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13833 * config/i386/sse.md (*mov<mode>_internal): Merge with
13834 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13835 Emit insn template depending on type attribute. Use
13836 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13837 movd instead of movq mnemonic for interunit moves. Rewrite mode
13838 attribute calculation. Remove unit attribute calculation.
13839 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13840 Set prefix_data16 attribute for DImode ssemov types.
13841 Use Ym instead of y for SSE-MMX conversion alternatives.
13842 Reorder operand constraints.
13844 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13846 * df.h (df_insn_delete): Adjust prototype.
13847 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13848 and let it decide whether mark the basic block dirty.
13849 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13850 * df-scan.c (df_insn_info_delete): New helper function, split
13851 off from df_insn_delete.
13852 (df_scan_free_bb_info): Use it.
13853 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13855 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13856 that the insn is actually an insn and it has a non-NULL basic block.
13857 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13859 2013-03-22 Richard Biener <rguenther@suse.de>
13861 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13863 (mem_ref_alloc): Do not allocate them.
13864 (refs_independent_p): Do not query or maintain a cache.
13866 2013-03-22 Richard Biener <rguenther@suse.de>
13868 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13869 (gather_mem_refs_in_loops): Do not compute it.
13870 (analyze_memory_references): Do not allocate it.
13871 (tree_ssa_lim_finalize): Do not free it.
13872 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13874 2013-03-22 Richard Biener <rguenther@suse.de>
13876 * is-a.h (as_a): Use gcc_checking_assert.
13878 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13880 * config/aarch64/aarch64.c (aarch64_print_operand): New
13881 format specifier for printing a constant in hex.
13882 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13883 format specifier for printing second operand.
13885 2013-03-22 Richard Biener <rguenther@suse.de>
13887 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13889 (gather_mem_refs_in_loops): Perform store accumulation here.
13890 (create_vop_ref_mapping_loop): Remove.
13891 (create_vop_ref_mapping): Likewise.
13892 (analyze_memory_references): Initialize refs_stored_in_loop.
13893 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13894 (record_indep_loop): Remove.
13895 (record_dep_loop): New function.
13896 (ref_indep_loop_p_1): Adjust to only walk over references
13897 in the loop, not its subloops.
13898 (ref_indep_loop_p): Rename to ...
13899 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13900 maintaining a more fine-grained cache.
13901 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13902 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13904 2013-03-22 Richard Biener <rguenther@suse.de>
13906 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13907 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13908 aggregate mem_ref_loc.
13909 (free_mem_ref_locs): Inline into ...
13910 (memref_free): ... this and adjust.
13911 (mem_ref_alloc): Adjust.
13912 (mem_ref_locs_alloc): Remove.
13913 (record_mem_ref_loc): Adjust.
13914 (get_all_locs_in_loop): Rewrite into ...
13915 (for_all_locs_in_loop): ... this iterator.
13916 (rewrite_mem_ref_loc): New functor.
13917 (rewrite_mem_refs): Use for_all_locs_in_loop.
13918 (sm_set_flag_if_changed): New functor.
13919 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13920 (ref_always_accessed): New functor.
13921 (ref_always_accessed_p): Use for_all_locs_in_loop.
13923 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13925 * tree-pass.h (PROP_gimple_lvec): New.
13926 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13927 (init_optimization_passes): Move pass_lower_vector.
13928 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13930 (pass_lower_vector): Provide PROP_gimple_lvec.
13931 (pass_lower_vector_ssa): Likewise.
13932 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13934 2013-03-21 Mark Wielaard <mjw@redhat.com>
13936 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13938 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13940 * config/i386/i386.md (*movdi_internal): Disparage slightly
13941 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13942 conversion alternatives.
13944 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13946 PR middle-end/48087
13947 * diagnostic.def (DK_WERROR): New kind.
13948 * diagnostic.h (werrorcount): Define.
13949 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13950 promoted to DK_ERROR, increment DK_WERROR counter instead of
13952 * toplev.c (toplev_main): Call print_ignored_options even if
13953 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13954 even if just werrorcount is non-zero.
13957 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13959 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13960 (string_cst_pool_decl): New function.
13961 (optimize_one_addr_into_implicit_ptr): New function.
13962 (resolve_addr_in_expr): Optimize DWARF location expression
13963 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13964 which doesn't live in memory, but has DW_AT_location or
13965 DW_AT_const_value, or refers to a string literal, into
13966 DW_OP_GNU_implicit_pointer.
13967 (optimize_location_into_implicit_ptr): New function.
13968 (resolve_addr): If removing DW_AT_location of a variable because
13969 it was DW_OP_addr of address of the variable, but the variable doesn't
13970 live in memory, try to emit const value attribute for the initializer.
13972 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13974 * tree.h (VECTOR_TYPE_P): New macro.
13975 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13976 TYPE_MODE): Use it.
13977 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13978 VEC_COND_EXPR cannot be lvalues.
13979 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13981 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13983 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13984 Restrict the transformation to equal modes.
13986 2013-03-21 Richard Biener <rguenther@suse.de>
13988 PR tree-optimization/39326
13989 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13990 (MEM_ANALYZABLE): Adjust.
13991 (record_mem_ref_loc): Move bitmap ops ...
13992 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13993 unanalyzable refs, do not record locations for it.
13994 (analyze_memory_references): Allocate ref zero as shared
13996 (refs_independent_p): Do not test for unanalyzed mems here.
13997 (ref_indep_loop_p_1): Special-case disambiguation against
13998 the unanalyzed ref.
13999 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14001 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14003 * config/arm/arm-protos.h (tune_params): Add
14004 prefer_neon_for_64bits field.
14005 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14006 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14007 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14008 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14009 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14010 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14011 (arm_option_override): Handle -mneon-for-64bits new option.
14012 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14013 (prefer_neon_for_64bits): Declare new variable.
14014 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14015 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14016 (arch_enabled): Handle new arch types. Remove support for onlya8
14018 (one_cmpldi2): Use new arch names.
14019 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14020 * config/arm/arm.opt (mneon-for-64bits): Add option.
14021 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14022 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14023 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14025 * doc/invoke.texi (-mneon-for-64bits): Document.
14027 2013-03-21 Richard Biener <rguenther@suse.de>
14029 PR tree-optimization/39326
14030 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14031 (sort_bbs_in_loop_postorder_cmp): New function.
14032 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14034 2013-03-21 Richard Biener <rguenther@suse.de>
14036 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14037 (vect_insert_into_interleaving_chain): Likewise.
14038 (vect_drs_dependent_in_basic_block): Inline ...
14039 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14041 (vect_analyze_data_ref_dependence): ... here. Simplify.
14042 (vect_check_interleaving): Simplify.
14043 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14044 (vect_slp_analyze_data_ref_dependences): ... this new function.
14045 (dr_group_sort_cmp): New function.
14046 (vect_analyze_data_ref_accesses): Compute data-reference groups
14047 here instead of in vect_analyze_data_ref_dependence. Use
14048 a more efficient algorithm.
14049 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14050 vect_slp_analyze_data_ref_dependences. Call
14051 vect_analyze_data_ref_accesses earlier.
14052 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14053 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14054 (vect_slp_analyze_data_ref_dependences): New prototype.
14056 2013-03-21 Richard Biener <rguenther@suse.de>
14058 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14059 ref is stored in the loop.
14060 (find_refs_for_sm): Walk only over all stores.
14061 (store_motion_loop): Allocate from lim_bitmap_obstack.
14062 (store_motion): Likewise.
14064 2013-03-21 Richard Biener <rguenther@suse.de>
14066 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14067 Update virtual SSA form.
14069 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14071 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14072 * configure: Regenerate.
14073 * config.in: Regenerate.
14074 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14075 if !HAVE_LD_EH_FRAME_CIEV3.
14077 2013-03-21 Richard Biener <rguenther@suse.de>
14079 * tree-cfg.c (verify_expr_no_block): New function.
14080 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14081 nor DECL_VALUE_EXPR have locations with associated blocks.
14082 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14083 (clear_unused_block_pointer): Remove code dealing with
14084 blocks in DECL_DEBUG_EXPR locations.
14086 2013-03-21 Richard Biener <rguenther@suse.de>
14088 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14089 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14090 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14091 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14092 instead of DECL_DEBUG_EXPR_IS_FROM.
14093 * gimplify.c (gimplify_modify_expr): Likewise.
14094 * tree-cfg.c (verify_expr_location_1): Likewise.
14095 * tree-complex.c (create_one_component_var): Likewise.
14096 * tree-sra.c (create_access_replacement): Likewise.
14097 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14098 (clear_unused_block_pointer): Likewise.
14099 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14100 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14101 * var-tracking.c (var_debug_decl): Likewise.
14102 (track_expr_p): Likewise.
14103 * tree-inline.c (add_local_variables): Likewise. Set
14104 DECL_HAS_DEBUG_EXPR_P after copying it.
14105 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14106 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14108 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14111 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14112 * configure: Regenerate.
14113 * config.in: Regenerate.
14114 * config/i386/i386.md (*movdf_internal): Use
14115 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14116 movd instead of movq mnemonic for interunit moves.
14117 (*movdi_internal): Ditto.
14119 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14121 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14122 (abd<mode>_3): New pattern.
14123 (aba<mode>_3): New pattern.
14124 (fabd<mode>_3): New pattern.
14126 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14128 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14129 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14130 occurrence of REGISTER_PREFIX as its empty string.
14132 2013-03-20 Jeff Law <law@redhat.com>
14134 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14135 addititional equivalences for equality comparisons between an SSA_NAME
14136 and a constant where the SSA_NAME was set from a widening conversion.
14138 2013-03-20 Walter Lee <walt@tilera.com>
14140 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14142 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14144 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14145 depending on type attribute.
14146 (*movti_internal): Ditto.
14147 (*movtf_internal): Ditto.
14148 (*movxf_internal): Ditto.
14149 (*movdf_internal): Ditto.
14150 (*movsf_internal): Ditto.
14152 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14154 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14155 maybe_vex for sselog1 and ssemov types.
14156 (*movdi_internal): Reorder operand constraints.
14157 (*movsi_internal): Ditto. Set prefix attribute to
14158 maybe_vex for sselog1 and ssemov types.
14159 (*movtf_internal): Set prefix attribute to maybe_vex
14160 for sselog1 and ssemov types.
14161 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14162 DImode ssemov types. Reorder operand constraints.
14163 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14164 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14165 attribute for SImode ssemov types. Reorder operand constraints.
14167 2013-03-20 Martin Jambor <mjambor@suse.cz>
14169 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14170 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14172 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14174 * config/rs6000/predicates.md (indexed_address, update_address_mem
14175 update_indexed_address_mem): New predicates.
14176 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14177 attribute for load/store instructions.
14178 * config/rs6000/dfp.md (movsd_store): Likewise.
14179 (movsd_load): Likewise.
14180 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14181 (unnamed HI->DI extend define_insn): Likewise.
14182 (unnamed SI->DI extend define_insn): Likewise.
14183 (unnamed QI->SI extend define_insn): Likewise.
14184 (unnamed QI->HI extend define_insn): Likewise.
14185 (unnamed HI->SI extend define_insn): Likewise.
14186 (unnamed HI->SI extend define_insn): Likewise.
14187 (extendsfdf2_fpr): Likewise.
14188 (movsi_internal1): Likewise.
14189 (movsi_internal1_single): Likewise.
14190 (movhi_internal): Likewise.
14191 (movqi_internal): Likewise.
14192 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14193 attribute for load/store instructions.
14194 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
14196 (mov<mode>_softfloat): Likewise.
14197 (mov<mode>_hardfloat32): Likewise.
14198 (mov<mode>_hardfloat64): Likewise.
14199 (mov<mode>_softfloat64): Likewise.
14200 (movdi_internal32): Likewise.
14201 (movdi_internal64): Likewise.
14202 (probe_stack_<mode>): Likewise.
14204 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
14206 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
14207 floating point, and decimal floating point to reload iterator.
14209 * config/rs6000/constraints.md (wl constraint): New constraints to
14210 return FLOAT_REGS if certain options are used to reduce the number
14211 of separate patterns that exist in the file.
14212 (wx constraint): Likewise.
14213 (wz constraint): Likewise.
14215 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14216 -mdebug=reg, print wg, wl, wx, and wz constraints.
14217 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14218 Initialize the reload functions for 64-bit binary/decimal floating
14220 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14221 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14222 create the buffer on the stack to overcome not having a 32-bit
14224 (rs6000_emit_move): Likewise.
14225 (rs6000_secondary_memory_needed_rtx): Likewise.
14226 (rs6000_alloc_sdmode_stack_slot): Likewise.
14227 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14228 via xxlxor, just like DFmode 0.0.
14230 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
14231 define as 1 if we are running on a power7 or newer.
14232 (enum r6000_reg_class_enum): Add new constraints.
14234 * config/rs6000/dfp.md (movsd): Delete, combine with binary
14235 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14236 with other moves by using conditional constraits (wg). Use LFIWZX
14237 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
14238 (movsd splitter): Likewise.
14239 (movsd_hardfloat): Likewise.
14240 (movsd_softfloat): Likewise.
14242 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
14243 binary and decimal floating point moves.
14244 (fmove_ok): New attributes to combine binary and decimal floating
14245 point moves, and to combine power6x (mfpgpr) moves along normal
14247 (real_value_to_target): Likewise.
14248 (f32_lr): Likewise.
14249 (f32_lm): Likewise.
14250 (f32_li): Likewise.
14251 (f32_sr): Likewise.
14252 (f32_sm): Likewise.
14253 (f32_si): Likewise.
14254 (movsf): Combine binary and decimal floating point moves. Combine
14255 power6x (mfpgpr) moves with other moves by using conditional
14256 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
14257 (mov<mode> for SFmode/SDmode); Likewise.
14258 (SFmode/SDmode splitters): Likewise.
14259 (movsf_hardfloat): Likewise.
14260 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
14261 (movsf_softfloat): Likewise.
14262 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
14264 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
14265 wx and wz constraints.
14267 * config/rs6000/constraints.md (wg constraint): New constraint to
14268 return FLOAT_REGS if -mmfpgpr (power6x) was used.
14270 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
14273 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14274 -mdebug=reg, print wg, wl, wx, and wz constraints.
14275 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14276 Initialize the reload functions for 64-bit binary/decimal floating
14278 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14279 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14280 create the buffer on the stack to overcome not having a 32-bit
14282 (rs6000_emit_move): Likewise.
14283 (rs6000_secondary_memory_needed_rtx): Likewise.
14284 (rs6000_alloc_sdmode_stack_slot): Likewise.
14285 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14286 via xxlxor, just like DFmode 0.0.
14288 * config/rs6000/dfp.md (movdd): Delete, combine with binary
14289 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14290 with other moves by using conditional constraits (wg). Use LFIWZX
14291 and STFIWX for loading SDmode on power7.
14292 (movdd splitters): Likewise.
14293 (movdd_hardfloat32): Likewise.
14294 (movdd_softfloat32): Likewise.
14295 (movdd_hardfloat64_mfpgpr): Likewise.
14296 (movdd_hardfloat64): Likewise.
14297 (movdd_softfloat64): Likewise.
14299 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
14300 64-bit binary and decimal floating point moves.
14301 (FMOVE64X): Likewise.
14302 (movdf): Combine 64-bit binary and decimal floating point moves.
14303 Combine power6x (mfpgpr) moves with other moves by using
14304 conditional constraits (wg).
14305 (mov<mode> for DFmode/DDmode): Likewise.
14306 (DFmode/DDmode splitters): Likewise.
14307 (movdf_hardfloat32): Likewise.
14308 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
14309 (movdf_softfloat32): Likewise.
14310 (movdf_hardfloat64_mfpgpr): Likewise.
14311 (movdf_hardfloat64): Likewise.
14312 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
14313 (movdf_softfloat64): Likewise.
14314 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
14315 (reload_<mode>_load): Move to later in the file so they aren't in
14316 the middle of the floating point move insns.
14317 (reload_<mode>_store): Likewise.
14319 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
14322 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
14323 constraint if -mdebug=reg.
14324 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
14325 Enable using dd reload support if needed.
14327 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
14328 binary and decimal floating point moves in rs6000.md.
14329 (movtd_internal): Likewise.
14331 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
14332 decimal floating point moves.
14334 (movtf_internal): Likewise.
14335 (mov<mode>_internal, TDmode/TFmode): Likewise.
14336 (movtf_softfloat): Likewise.
14337 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
14339 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
14340 movdi_internal64, using wg constraint for move direct operations.
14341 (movdi_internal64): Likewise.
14343 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
14344 MODES_TIEABLE_P for selected modes. Print the numerical value of
14345 the various virtual registers. Use GPR/FPR first/last values,
14346 instead of hard coding the register numbers. Print which modes
14347 have reload functions registered.
14348 (rs6000_option_override_internal): If -mdebug=reg, trace the options
14349 settings before/after setting cpu, target and subtarget settings.
14350 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
14351 and for secondary reload failures in rs6000_secondary_reload_inner.
14352 (rs6000_secondary_reload_fail): Likewise.
14353 (rs6000_secondary_reload_inner): Likewise.
14355 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
14356 macros for first/last GPR and FPR registers.
14357 (LAST_GPR_REGNO): Likewise.
14358 (FIRST_FPR_REGNO): Likewise.
14359 (LAST_FPR_REGNO): Likewise.
14361 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
14362 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
14363 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
14364 (vcond<mode><mode>): Likewise.
14365 (vcondu<mode><mode>): Likewise.
14366 (vector_gtu<mode>): Likewise.
14367 (vector_gte<mode>): Likewise.
14368 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
14369 to prevent the compiler from converting DImode operations to TImode.
14370 (ior<mode>3): Likewise.
14371 (and<mode>3): Likewise.
14372 (one_cmpl<mode>2): Likewise.
14373 (nor<mode>3): Likewise.
14374 (andc<mode>3): Likewise.
14376 * config/rs6000/constraints.md (wt constraint): New constraint
14377 that returns VSX_REGS if TImode is allowed in VSX registers.
14379 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
14380 constant under VSX.
14382 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
14383 similar to TImode, but it is restricted to being in the GPRs.
14385 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
14386 TImode to occupy a single VSX register.
14388 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
14389 -mvsx-timode for power7/power8.
14390 (power7 cpu): Likewise.
14391 (power8 cpu): Likewise.
14393 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
14394 sure that TFmode/TDmode take up two registers if they are ever
14395 allowed in the upper VSX registers.
14396 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
14398 (rs6000_init_hard_regno_mode_ok): Likewise.
14399 (rs6000_debug_reg_global): Add debugging for PTImode and wt
14400 constraint. Print if LRA is turned on.
14401 (rs6000_option_override_internal): Give an error if -mvsx-timode
14402 and VSX is not enabled.
14403 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
14404 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
14405 to reg+offset addressing. Use PTImode when checking offset
14406 addresses for validity.
14407 (reg_offset_addressing_ok_p): Likewise.
14408 (rs6000_legitimate_offset_address_p): Likewise.
14409 (rs6000_legitimize_address): Likewise.
14410 (rs6000_legitimize_reload_address): Likewise.
14411 (rs6000_legitimate_address_p): Likewise.
14412 (rs6000_eliminate_indexed_memrefs): Likewise.
14413 (rs6000_emit_move): Likewise.
14414 (rs6000_secondary_reload): Likewise.
14415 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
14416 reloads to fpr registers to continue to use reg+offset addressing,
14417 but 64-bit reloads to altivec registers need reg+reg addressing.
14418 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
14419 it. Treat LO_SUM like a PLUS operation.
14420 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
14421 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
14422 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
14423 registers to share a register with a smaller sized type, since VSX
14424 puts scalars in the upper 64-bits.
14425 (print_operand): Add support for PTImode.
14426 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
14427 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
14428 registers, but don't have arithmetic support.
14429 (rs6000_memory_move_cost): Add test for VSX.
14430 (rs6000_opt_masks): Add -mvsx-timode.
14432 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
14435 (VSr): Use wt constraint for TImode.
14436 (VSv): Drop TImode support.
14437 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
14438 (vsx_movti_64bit): Likewise.
14439 (vsx_movti_32bit): Likewise.
14440 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
14441 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
14442 one '?' on the appropriate output constraint. Do not allow TImode
14443 logical operations on 32-bit systems.
14444 (vsx_ior<mode>3): Likewise.
14445 (vsx_xor<mode>3): Likewise.
14446 (vsx_one_cmpl<mode>2): Likewise.
14447 (vsx_nor<mode>3): Likewise.
14448 (vsx_andc<mode>3): Likewise.
14449 (vsx_concat_<mode>): Likewise.
14450 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
14452 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
14453 OPTION_MASK_VSX_TIMODE.
14454 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
14455 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
14457 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
14458 (TI2 iterator): New iterator for TImode, PTImode.
14459 (wd mode attribute): Add values for vector types.
14460 (movti_string): Replace TI move operations with operations for TImode
14461 and PTImode. Add support for TImode being allowed in VSX registers.
14462 (mov<mode>_string, TImode/PTImode): Likewise.
14463 (movti_ppc64): Likewise.
14464 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14465 (TI mode splitters): Likewise.
14467 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14470 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14472 PR tree-optimization/56355
14473 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14474 Also handle integers with undefined overflow.
14476 2013-03-20 Catherine Moore <clm@codesourcery.com>
14477 Maciej W. Rozycki <macro@codesourcery.com>
14478 Tom de Vries <tom@codesourcery.com>
14479 Nathan Sidwell <nathan@codesourcery.com>
14480 Iain Sandoe <iain@codesourcery.com>
14481 Nathan Froyd <froydnj@codesourcery.com>
14482 Chao-ying Fu <fu@mips.com>
14484 * doc/extend.texi: (micromips, nomicromips, nocompression):
14485 Document new function attributes.
14486 * doc/invoke.texi (minterlink-compressed, mmicromips,
14487 m14k, m14ke, m14kec): Document new options.
14488 (minterlink-mips16): Update documentation.
14489 * doc/md.texi (ZC, ZD): Document new constraints.
14490 * configure.ac (gcc_cv_as_micromips): Check if linker
14491 supports the .set micromips directive.
14492 * configure: Regenerate.
14493 * config.in: Regenerate.
14494 * config/mips/mips-tables.opt: Regenerate.
14495 * config/mips/micromips.md: New file.
14496 * constraints.md (ZC, ZD): New constraints.
14497 * config/mips/predicates.md (movep_src_register): New predicate.
14498 (movep_src_operand): New predicate.
14499 (non_volatile_mem_operand): New predicate.
14500 * config/mips/mips.md (multimem): New type.
14501 (length): Differentiate between 17-bit and 18-bit branch offsets.
14502 (MOVEP1, MOVEP2): New mode iterator.
14503 (mov_<load>l): Use ZC constraint.
14504 (mov_<load>r): Likewise.
14505 (mov_<store>l): Likewise.
14506 (mov_<store>r): Likewise.
14507 (*branch_equality<mode>_inverted): Add microMIPS support.
14508 (*branch_equality<mode>): Likewise.
14509 (*jump_absolute): Likewise.
14510 (indirect_jump_<mode>): Likewise.
14511 (tablejump_<mode>): Likewise.
14512 (<optab>_internal): Likewise.
14513 (sibcall_internal): Likewise.
14514 (sibcall_value_internal): Likewise.
14515 (prefetch): Use constraint ZD.
14516 * config/mips/mips.opt (minterlink-compressed): New option.
14517 (minterlink-mips16): Now an alias for minterlink-compressed.
14518 (mmicromips): New option.
14519 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14520 (compare_and_swap_12): Likewise.
14521 (sync_add<mode>): Likewise.
14522 (sync_<optab>_12): Likewise.
14523 (sync_old_<optab>_12): Likewise.
14524 (sync_new_<optab>_12): Likewise.
14525 (sync_nand_12): Likewise.
14526 (sync_old_nand_12): Likewise.
14527 (sync_new_nand_12): Likewise.
14528 (sync_sub<mode>): Likewise.
14529 (sync_old_add<mode>): Likewise.
14530 (sync_old_sub<mode>): Likewise.
14531 (sync_new_add<mode>): Likewise.
14532 (sync_new_sub<mode>): Likewise.
14533 (sync_<optab><mode>): Likewise.
14534 (sync_old_<optab><mode>): Likewise.
14535 (sync_new_<optab><mode>): Likewise.
14536 (sync_nand<mode>): Likewise.
14537 (sync_old_nand<mode>): Likewise.
14538 (sync_new_nand<mode>): Likewise.
14539 (sync_lock_test_and_set<mode>): Likewise.
14540 (test_and_set_12): Likewise.
14541 (atomic_compare_and_swap<mode>): Likewise.
14542 (atomic_exchange<mode>_llsc): Likewise.
14543 (atomic_fetch_add<mode>_llsc): Likewise.
14544 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14545 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14546 (umips_save_restore_pattern_p): Likewise.
14547 (umips_load_store_pair_p): Likewise.
14548 (umips_output_load_store_pair): Likewise.
14549 (umips_movep_target_p): Likewise.
14550 (umips_12bit_offset_address_p): Likewise.
14551 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14552 (mips_base_mips16): Rename this...
14553 (mips_base_compression_flags): ...to this. Update all uses.
14554 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14555 (mips_mips16_decl_p): Delete.
14556 (mips_nomips16_decl_p): Delete.
14557 (mips_get_compress_on_flags): New function.
14558 (mips_get_compress_off_flags): New function.
14559 (mips_get_compress_mode): New function.
14560 (mips_get_compress_on_name): New function.
14561 (mips_get_compress_off_name): New function.
14562 (mips_insert_attributes): Support multiple compression types.
14563 (mips_merge_decl_attributes): Likewise.
14564 (umips_12bit_offset_address_p): New function.
14565 (mips_start_function_definition): Emit .set micromips directive.
14566 (mips_call_may_need_jalx_p): New function.
14567 (mips_function_ok_for_sibcall): Add microMIPS support.
14568 (mips_print_operand_punctuation): Support short delay slots and
14570 (umips_swm_mask, umips_swm_encoding): New.
14571 (umips_build_save_restore): New function.
14572 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14573 (was_mips16_p): Remove.
14574 (old_compression_mode): New.
14575 (mips_set_compression_mode): New function.
14576 (mips_set_current_function): Add microMIPS support.
14577 (mips_option_override): Likewise.
14578 (umips_save_restore_pattern_p): New function.
14579 (umips_output_save_restore): New function.
14580 (umips_load_store_pair_p_1): New function.
14581 (umips_load_store_pair_p): New function.
14582 (umips_output_load_store_pair_1): New function.
14583 (umips_output_load_store_pair): New function.
14584 (umips_movep_target_p) New function.
14585 (mips_prepare_pch_save): Add microMIPS support.
14586 * config/mips/mips.h (TARGET_COMPRESSION): New.
14587 (TARGET_CPU_CPP_BUILTINS): Update macro
14588 to use new compression flags and to support microMIPS.
14589 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14590 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14591 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14592 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14593 (ASM_SPEC): Support mmicromips and mno-micromips.
14594 (M16STORE_REG_P): New macro.
14595 (MIPS_CALL): Support TARGET_MICROMIPS.
14596 (MICROMIPS_J): New macro.
14597 (mips_base_mips16): Rename this...
14598 (mips_base_compression_flags): ...to this.
14599 (UMIPS_12BIT_OFFSET_P): New macro.
14600 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14601 (MULTILIB_DIRNAMES): Likewise.
14602 2013-03-20 Richard Biener <rguenther@suse.de>
14604 PR tree-optimization/56661
14605 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14606 the result does not have to be distinct.
14608 2013-03-20 Richard Biener <rguenther@suse.de>
14610 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14613 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14614 Steven Bosscher <steven@gcc.gnu.org>
14616 PR rtl-optimization/56605
14617 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14619 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14622 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14623 that require movd instead of movq.
14625 2013-03-20 Richard Biener <rguenther@suse.de>
14627 * tree-ssa-structalias.c (struct variable_info): Add pointer
14628 to the first field of an aggregate with sub-vars. Make
14629 this and the pointer to the next subfield its ID.
14630 (vi_next): New function.
14631 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14632 storedanything_id, integer_id): Increment by one.
14633 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14634 get_call_clobber_vi): Adjust.
14635 (solution_set_expand): Simplify and speedup.
14636 (solution_set_add): Inline into ...
14637 (set_union_with_increment): ... this. Adjust accordingly.
14638 (do_sd_constraint): Likewise.
14639 (do_ds_constraint): Likewise.
14640 (do_complex_constraint): Simplify.
14641 (build_pred_graph): Adjust.
14642 (solve_graph): Likewise. Simplify and speedup.
14643 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14644 get_constraint_for_component_ref, get_constraint_for_1,
14645 first_vi_for_offset, first_or_preceding_vi_for_offset,
14646 create_function_info_for, create_variable_info_for_1,
14647 create_variable_info_for, intra_create_variable_infos): Adjust.
14648 (init_base_vars): Push NULL for ID zero.
14649 (compute_points_to_sets): Adjust.
14651 2013-03-20 Richard Biener <rguenther@suse.de>
14653 * cfgloop.c (verify_loop_structure): Streamline and avoid
14654 ICEing on corrupt loop tree.
14655 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14658 2013-03-20 Richard Biener <rguenther@suse.de>
14660 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14661 check whether an SSA update is needed.
14663 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14665 * config/mips/constraints.md (T): Rename to...
14669 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14670 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14672 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14674 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14675 (*subsi3_carryin_uxtw): Likewise.
14677 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14679 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14680 (*rorsi3_insn_uxtw): Likewise.
14682 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14684 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14685 (*extrsi5_insn_uxtw): Likewise.
14687 2013-03-19 Richard Biener <rguenther@suse.de>
14689 PR tree-optimization/56273
14690 * passes.c (init_optimization_passes): Move second VRP after DOM.
14692 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14694 * config/i386/i386.md (*movti_internal): Merge from
14695 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14696 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14697 nox64 isa attributes.
14699 2013-03-18 Richard Biener <rguenther@suse.de>
14701 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14703 (merge_node_constraints): Likewise.
14704 (build_succ_graph): Likewise.
14705 (valid_graph_edge): Inline into single caller.
14706 (unify_nodes): Likewise. Use bitmap_set_bit return value
14708 (scc_visit): Fix formatting and variable use.
14709 (do_sd_constraint): Use gcc_checking_assert.
14710 (do_ds_constraint): Likewise.
14711 (do_complex_constraint): Likewise.
14712 (condense_visit): Likewise. Cleanup.
14713 (dump_pred_graph): New function.
14714 (perform_var_substitution): Dump the pred-graph before
14715 variable substitution.
14716 (find_equivalent_node): Use gcc_checking_assert.
14717 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14719 2013-03-18 Richard Biener <rguenther@suse.de>
14721 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14722 Remove cond_expr_stmt_list argument and do not gimplify the
14724 (vect_loop_versioning): Adjust.
14725 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14726 Cleanup to use less temporaries.
14727 (vect_create_data_ref_ptr): Cleanup.
14729 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14731 PR tree-optimization/56635
14732 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14733 require types_compatible_p types.
14735 2013-03-18 Nick Clifton <nickc@redhat.com>
14737 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14738 spurious backslash.
14740 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14741 Add missing line to comment describing function.
14743 2013-03-18 Richard Biener <rguenther@suse.de>
14745 PR tree-optimization/56210
14746 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14747 Handle string / character search functions.
14748 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14750 2013-03-18 Richard Biener <rguenther@suse.de>
14752 PR middle-end/56483
14753 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14754 and implement properly.
14755 * gimple.h (gimple_cond_single_var_p): Remove.
14757 2013-03-18 Richard Biener <rguenther@suse.de>
14759 * tree-data-ref.h (find_data_references_in_loop): Declare.
14760 * tree-data-ref.c (get_references_in_stmt): Use a stack
14761 vector pre-allocated in the callers.
14762 (find_data_references_in_stmt): Adjust.
14763 (graphite_find_data_references_in_stmt): Likewise.
14764 (create_rdg_vertices): Likewise.
14765 (find_data_references_in_loop): Export.
14766 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14767 Compute dependences here...
14768 (vect_analyze_data_refs): ...not here. When we encounter
14769 a non-vectorizable data reference in basic-block vectorization
14770 truncate the data reference vector. Do not bother to
14771 fixup data-dependence information for gather loads.
14772 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14773 of data references, as reported.
14775 2013-03-18 Richard Biener <rguenther@suse.de>
14777 PR tree-optimization/3713
14778 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14779 has_constants and expr.
14780 (stmt_has_constants): Properly valueize SSA names when deciding
14781 whether the stmt has constants.
14783 2013-03-18 Richard Biener <rguenther@suse.de>
14785 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14786 whole function when there is nothing to do.
14787 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14788 * tree-vectorizer.c (vectorize_loops): Update virtual and
14789 loop-closed SSA once.
14790 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14792 2013-03-18 Richard Biener <rguenther@suse.de>
14794 PR middle-end/56113
14795 * domwalk.c (bb_postorder): New global static.
14796 (cmp_bb_postorder): New function.
14797 (walk_dominator_tree): Replace scheme imposing an order for
14798 visiting dominator sons by one sorting them at the time they
14799 are pushed on the stack.
14801 2013-03-18 Richard Biener <rguenther@suse.de>
14803 PR tree-optimization/39326
14804 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14805 (struct mem_ref): Replace mem member with ao_ref typed member.
14806 (MEM_ANALYZABLE): Adjust.
14807 (memref_eq): Likewise.
14808 (mem_ref_alloc): Likewise.
14809 (gather_mem_refs_stmt): Likewise.
14810 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14811 (execute_sm_if_changed_flag_set): Adjust.
14812 (execute_sm): Likewise.
14813 (ref_always_accessed_p): Likewise.
14814 (refs_independent_p): Likewise.
14815 (can_sm_ref_p): Likewise.
14817 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14820 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14821 return 1 even for !unsignedp.
14823 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14825 * config/i386/i386.md (isa): Add x64 and nox64.
14826 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14827 (*pushtf): Enable *roF alternative for x64 isa only.
14828 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14829 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14830 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14831 (*movtf_internal): Merge from *movtf_internal_rex64 and
14832 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14833 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14834 nox64 isa attributes.
14835 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14836 nox64 isa attributes.
14837 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14839 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14841 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14843 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14846 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14847 class names. Remove trailing comma after "ALL_REGS".
14849 2013-03-16 Jan Hubicka <jh@suse.cz>
14851 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14852 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14853 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14854 of cgraph_get_create_node.
14855 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14857 2013-03-16 Jason Merrill <jason@redhat.com>
14860 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14861 with DW_AT_default_value.
14863 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14865 * BASE-VER: Set to 4.9.0.
14867 2013-03-14 Andi Kleen <ak@linux.intel.com>
14870 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14871 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14872 Document _x* TSX intrinsics.
14874 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14875 David Holsgrove <david.holsgrove@xilinx.com>
14877 * configure.ac: Add MicroBlaze TLS support detection.
14878 * configure: Regenerate.
14879 * config/microblaze/microblaze-protos.h
14880 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14881 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14882 * config/microblaze/microblaze.c (microblaze_address_type): Add
14883 ADDRESS_TLS and tls_reloc address types.
14884 (microblaze_address_info): Add tls_reloc.
14885 (TARGET_HAVE_TLS): Define.
14886 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14887 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14888 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14889 load_tls_operand, microblaze_call_tls_get_addr,
14890 microblaze_legitimize_tls_address): New functions.
14891 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14892 (get_base_reg): Use microblaze_tls_symbol_p.
14893 (microblaze_classify_address): Handle TLS.
14894 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14895 label_mentioned_p and microblaze_tls_referenced_p.
14896 (microblaze_legitimize_address): Handle TLS.
14897 (microblaze_address_insns): Handle ADDRESS_TLS.
14898 (pic_address_needs_scratch): Handle TLS.
14899 (print_operand_address): Handle TLS.
14900 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14901 (microblaze_expand_move): Handle TLS.
14902 (microblaze_legitimate_constant_p): Check
14903 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14904 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14905 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14906 (PIC_OFFSET_TABLE_REGNUM): Set.
14907 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14908 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14909 (addsi3, movsi_internal2, movdf_internal): Update constraints
14910 * config/microblaze/predicates.md (arith_plus_operand): Define
14911 (move_operand): Redefine as move_src_operand,
14912 check microblaze_tls_referenced_p.
14914 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14916 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14917 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14919 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14921 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14924 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14926 PR tree-optimization/53265
14927 * common.opt (Waggressive-loop-optimizations): New option.
14928 * tree-ssa-loop-niter.c: Include tree-pass.h.
14929 (do_warn_aggressive_loop_optimizations): New function.
14930 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14931 if number_of_latch_executions returned constant.
14932 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14933 early. If number_of_latch_executions returned constant, set
14934 nb_iterations_upper_bound back to it.
14935 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14937 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14938 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14940 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14941 (MULTILIB_OSDIRNAMES): Set.
14942 * genmultilib: If defaultosdirname doesn't start with :: , set
14943 defaultosdirname2 instead, clear it and emit two . multilib_raw
14944 entries instead of just one.
14946 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14948 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14949 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14950 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14951 (SUBTARGET_OVERRIDE_OPTIONS): New.
14953 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14956 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14957 (musermode): Convert to Var(TARGET_USERMODE).
14958 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14959 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14960 * config/sh/sh.c (sh_option_override): Use
14961 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14962 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14964 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14966 (udivsi3_i4_single, divsi3_i4_single): Use
14967 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14969 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14971 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14974 2013-03-13 Richard Biener <rguenther@suse.de>
14976 PR tree-optimization/56608
14977 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14978 calls when vectorizing basic-blocks.
14980 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14983 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14986 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14988 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14990 2013-03-11 Jan Hubicka <jh@suse.cz>
14993 * lto-streamer-out.c (output_symbol_p): Skip references from
14994 constructors of external variables.
14996 2013-03-11 Jan Hubicka <jh@suse.cz>
14998 PR middle-end/56571
14999 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15001 * emit-rtl.c (verify_rtx_sharing): Likewise.
15002 (copy_insn_1): Likewise.
15003 * rtl.c (copy_rtx): Likewise.
15005 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15008 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15009 output_operand_lossage message.
15011 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15014 * arm.c (shift_op): Validate RTL pattern on the fly.
15015 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15018 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15021 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15022 new call patterns that clobber %fr12.
15023 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15024 split and postreload patterns.
15025 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15026 registers %fr12 and %fr12R as call used.
15028 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15030 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15031 (canon_address, record_store, replace_read, check_mem_read_rtx,
15032 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15033 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15034 rest_of_handle_dse): Likewise.
15036 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15038 PR middle-end/56524
15039 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15041 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15042 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15043 (save_optabs_if_changed): Replace with...
15044 (init_tree_optimization_optabs): ...this.
15045 * optabs.c (save_optabs_if_changed): Rename to...
15046 (init_tree_optimization_optabs): ...this. Take the optimization node
15047 as argument. Do nothing if the base optabs are already correct.
15048 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15049 to recompute optabs.
15050 * function.h (function): Remove optabs field.
15051 * function.c (invoke_set_current_function_hook): Call
15052 init_tree_optimization_optabs. Use the result to initialize
15055 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15057 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15058 if GTMA_HAS_NO_INSTRUMENTATION.
15059 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15060 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15061 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15062 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15063 GTMA_HAS_NO_INSTRUMENTATION.
15065 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15067 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15070 2013-03-08 Marek Polacek <polacek@redhat.com>
15071 Jakub Jelinek <jakub@redhat.com>
15073 PR tree-optimization/56478
15074 * predict.c (is_comparison_with_loop_invariant_p): Change the
15075 type of loop_step to tree.
15076 (predict_loops): Adjust.
15077 (predict_iv_comparison): Perform the computations on double_ints.
15079 2013-03-08 Richard Biener <rguenther@suse.de>
15081 PR tree-optimization/56570
15082 * tree-cfg.c (verify_expr_location_1): Verify locations for
15084 * tree-sra.c (create_access_replacement): Strip locations
15085 from DECL_DEBUG_EXPRs.
15087 2013-03-08 Richard Biener <rguenther@suse.de>
15089 * tree-inline.c (expand_call_inline): Do not associate
15090 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15091 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15093 2013-03-08 Richard Biener <rguenther@suse.de>
15095 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15096 or block changes with -Og. Fix for location / block encoding
15097 changes and PHI arguments with locations.
15099 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15101 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15103 (struct output_info): Likewise.
15104 (register_overhead): Remove bad gcc_assert.
15105 (bitmap_find_bit): If there is only a single bitmap element, do not
15106 count a miss as a search.
15107 (print_statistics): Update for counter type changes.
15108 (dump_bitmap_statistics): Likewise. Print headers such that they
15109 are properly lined up with the printed counters.
15111 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15113 PR tree-optimization/56559
15114 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15115 check that it has only a single use.
15117 2013-03-07 Richard Biener <rguenther@suse.de>
15119 * doc/invoke.texi (fwhole-program): Discourage use in combination
15122 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15124 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15126 PR tree-optimization/56539
15127 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15128 instead of GSI_CONTINUE_LINKING as last argument to
15129 force_gimple_operand_gsi. Adjust function comment.
15131 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15134 PR middle-end/56548
15135 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15136 promoted mode, convert the result back to the original mode.
15138 2013-03-06 Richard Biener <rguenther@suse.de>
15140 PR middle-end/56294
15141 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15142 (insert_updated_phi_nodes_compare_uids): New function.
15143 (update_ssa): Sort symbols_to_rename after UID before
15144 traversing it to insert PHI nodes.
15146 2013-03-06 Richard Biener <rguenther@suse.de>
15148 PR middle-end/50494
15149 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15150 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15153 2013-02-13 Richard Biener <rguenther@suse.de>
15156 * varasm.c (output_constant_def_1): Get the decl representing
15157 the constant as argument.
15158 (output_constant_def): Wrap output_constant_def_1.
15159 (make_decl_rtl): Use output_constant_def_1 with the decl
15160 representing the constant.
15161 (build_constant_desc): Optionally re-use a decl already
15162 representing the constant.
15163 (tree_output_constant_def): Adjust.
15165 2013-03-06 Joey Ye <joey.ye@arm.com>
15168 * gcc.c (convert_white_space): New function.
15169 (main): Handles white space in function name.
15171 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15174 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15175 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15176 to SH_DIV_CALL_TABLE for TARGET_SH2.
15177 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15179 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15180 call-table options.
15182 2013-03-05 Sterling Augustine <saugustine@google.com>
15183 Cary Coutant <ccoutant@google.com>
15186 * dwarf2out.c (resolve_addr): Don't call
15187 remove_loc_list_addr_table_entries a second time for the same
15190 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15193 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15194 (avoid_complex_debug_insns): New function.
15195 (expand_debug_locations): Call it.
15197 PR rtl-optimization/56484
15198 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
15199 lifetimes of hard registers on small register class machines.
15201 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15203 * config/microblaze/microblaze-protos.h: Rename
15204 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
15205 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
15207 (microblaze_fast_interrupt_function_p): New function.
15208 (microblaze_is_interrupt_handler): Rename to
15209 microblaze_is_interrupt_variant and add fast_interrupt check.
15210 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
15211 (save_restore_insns): Likewise.
15212 (compute_frame_size): Likewise.
15213 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
15214 (microblaze_globalize_label): Likewise.
15215 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
15216 * config/microblaze/microblaze.md: Use wrapper
15217 microblaze_is_interrupt_variant.
15219 2013-03-05 Kai Tietz <ktietz@redhat.com>
15221 * sdbout.c (sdbout_one_type): Switch to current function's section
15222 supporting cold/hot.
15224 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15226 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
15229 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15231 PR middle-end/56461
15232 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
15233 if VALGRIND_GET_VBITS is defined, temporarily make object
15234 memory all defined, and restore previous valgrind addressability
15235 and definability afterwards. Free this_object at the end.
15237 PR middle-end/56461
15238 * lra.c (lra): Call lra_clear_live_ranges if live_p,
15239 right before calling lra_create_live_ranges, also call it
15240 when clearing live_p. Only call lra_clear_live_ranges
15241 at the end if live_p.
15243 PR middle-end/56461
15244 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
15246 2013-03-05 Richard Biener <rguenther@suse.de>
15248 PR tree-optimization/56521
15249 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
15252 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
15255 * except.h (remove_unreachable_eh_regions): New prototype.
15256 * except.c (remove_eh_handler_splicer): New function, split out
15257 of remove_eh_handler.
15258 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
15259 warning about running it on many EH regions one at a time.
15260 (remove_unreachable_eh_regions_worker): New function, walk the
15261 EH tree in depth-first order and remove non-marked regions.
15262 (remove_unreachable_eh_regions): New function.
15263 * tree-eh.c (mark_reachable_handlers): New function, split out
15264 from remove_unreachable_handlers.
15265 (remove_unreachable_handlers): Use mark_reachable_handlers and
15266 remove_unreachable_eh_regions.
15267 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
15268 and remove_unreachable_eh_regions.
15270 2013-03-05 Richard Biener <rguenther@suse.de>
15272 PR middle-end/56525
15273 * loop-init.c (fix_loop_structure): Remove loops in two stages,
15274 not freeing them until the end.
15276 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15278 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
15280 2013-03-05 Richard Biener <rguenther@suse.de>
15282 PR tree-optimization/56270
15283 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
15284 of loads after scheduling an SLP instance.
15286 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15288 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
15290 (check_gcc_parallelize): Run guality.exp as a separate job from
15291 vect.exp with unsorted.exp and $(dg_target_exps) separately from
15292 struct-layout-1.exp with stackalign.exp.
15294 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
15296 PR middle-end/56461
15297 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
15298 load_index sbitmap even if some bit in it isn't set.
15300 PR middle-end/56461
15301 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
15302 (discover_iteration_bound_by_body_walk): Change queues to
15303 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
15304 spelling in comment. Call safe_push on queues[bound_index] directly.
15305 Release queues[queue_index] in every iteration unconditionally.
15306 Release bounds vector.
15308 PR middle-end/56461
15309 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
15310 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
15311 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
15314 2013-03-05 Richard Biener <rguenther@suse.de>
15317 * tree-inline.c (remap_blocks_to_null): New function.
15318 (expand_call_inline): When expanding a call stmt without
15319 an associated block inline remap all callee blocks to NULL.
15321 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15323 PR rtl-optimization/56494
15324 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
15325 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
15326 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
15328 PR middle-end/56461
15329 * sel-sched-ir.c (free_sched_pools): Release
15330 succs_info_pool.stack[succs_info_pool.max_top] vectors too
15331 if succs_info_pool.max_top isn't -1.
15334 * opts.c (opts_obstack, opts_concat): Moved to...
15335 * opts-common.c (opts_obstack, opts_concat): ... here.
15337 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15339 PR middle-end/56461
15340 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
15342 2013-03-04 Martin Jambor <mjambor@suse.cz>
15344 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
15345 all appropriate places.
15347 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
15349 PR tree-optimization/56424
15350 * ipa-split.c (split_function): Do not set the RSO flag if result is
15351 not by reference and its type is a register type.
15353 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
15355 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
15356 (microblaze_legitimate_pic_operand): Likewise
15357 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
15358 new function microblaze_legitimate_pic_operand
15359 * config/microblaze/microblaze-protos.h
15360 (microblaze_legitimate_pic_operand): Declare.
15362 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15364 * config/microblaze/predicates.md (call_insn_simple_operand):
15365 New predicate for supported rtx code types.
15366 * config/microblaze/microblaze.md (call_internal1): Use
15367 call_insn_simple_operand predicate.
15369 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15371 PR middle-end/56461
15372 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
15373 partitions.ordered_remove.
15375 PR middle-end/56461
15376 * tree-vect-stmts.c (vectorizable_conversion): Don't call
15377 vec_oprnds0.create (1) for modifier == NONE.
15379 PR middle-end/56461
15380 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
15381 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
15382 vec_oprnds1 right before pushing anything to it for
15385 PR middle-end/56461
15386 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
15387 set nbbs to 0 instead of having separate code path.
15388 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
15389 instead of false as last argument if returning NULL.
15391 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
15393 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
15394 the attribute is now called "target" instead of "option".
15395 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
15396 * doc/tm.texi.in (Target Attributes): Likewise document the correct
15397 attribute/pragma name for TARGET_OPTION_VALID_P and
15398 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
15399 * doc/tm.texi: Regenerated.
15401 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
15403 * config/microblaze/microblaze.c:
15404 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
15405 * config/microblaze/microblaze.h: Add -mxl-reorder to
15407 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
15408 instructions emitted if TARGET_REORDER.
15409 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
15410 or 0 for -m/-mno case, but initialises as 2 to detect default use case
15413 2013-03-01 Xinliang David Li <davidxl@google.com>
15415 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
15418 2013-03-01 Jakub Jelinek <jakub@redhat.com>
15420 PR middle-end/56461
15421 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
15422 vector even when returning true. Fix up function comment formatting.
15424 PR middle-end/56461
15425 * ira-build.c (ira_loop_nodes_count): New variable.
15426 (create_loop_tree_nodes): Initialize it.
15427 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
15429 PR middle-end/56461
15430 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
15431 method on dr_chain and result_chain.
15432 * tree-vect-stmts.c (vectorizable_store): Only call
15433 result_chain.create if j == 0.
15435 PR middle-end/56461
15436 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
15437 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
15438 before overwriting it.
15440 2013-03-01 Tobias Burnus <burnus@net-b.de>
15442 * doc/extended.texi (C Extensions): Change order in @menu
15444 (Other MIPS Built-in Functions): Move last MIPS entry before
15445 "picoChip Built-in Functions".
15446 (SH Built-in Functions): Move after RX Built-in Functions.
15447 * doc/gcc.texi (Introduction): Change order in @menu
15449 * doc/md.texi (Constraints): Ditto.
15450 * gty.texi (Type Information): Ditto.
15451 (User-provided marking routines for template types): Make
15453 * doc/invoke.texi (AArch64 Options): Move before
15454 "Adapteva Epiphany Options".
15456 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
15457 Jakub Jelinek <jakub@redhat.com>
15460 * asan.c (gate_asan): Lookup no_sanitize_address instead of
15461 no_address_safety_analysis attribute.
15462 * doc/extend.texi (no_address_safety_attribute): Rename to
15463 no_sanitize_address attribute, mention no_address_safety_analysis
15464 attribute as deprecated alias.
15466 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15468 PR middle-end/56461
15469 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15470 type to vec<vec<tree> > *.
15471 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15472 to be vec<tree> instead of vec<tree> *, set vec_defs
15473 to vNULL and call vec_defs.create (number_of_vects), adjust other
15475 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15476 vectorizable_condition): Adjust vect_get_slp_defs callers.
15478 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15480 * config/aarch64/aarch64.c
15481 (aarch64_float_const_representable): Remove unused variable.
15483 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15485 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15487 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15489 * config/aarch64/aarch64-builtins.c
15490 (aarch64_init_simd_builtins): Make static.
15492 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15494 * config/aarch64/aarch64.c
15495 (aarch64_simd_make_constant): Make static.
15497 2013-02-28 Martin Jambor <mjambor@suse.cz>
15499 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15500 with no initialization to the RHS of debug statements.
15502 2013-02-28 Martin Jambor <mjambor@suse.cz>
15504 PR tree-optimization/56294
15505 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15507 (get_access_replacement): Do not call create_access_replacement.
15508 Assert a replacement exists.
15509 (get_repl_default_def_ssa_name): Create the replacement declaration
15512 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15514 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15515 final_end_function.
15517 2013-02-28 Marek Polacek <polacek@redhat.com>
15519 PR rtl-optimization/56466
15520 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15521 if we're changing a loop.
15522 (peel_loops_completely): Likewise.
15524 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15527 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15529 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15532 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15533 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15534 INTX_FTYPE_FX, FX_FTYPE_INTX.
15535 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15537 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15539 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15540 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15541 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15542 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15543 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15544 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15545 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15546 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15547 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15548 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15549 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15550 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15551 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15552 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15553 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15554 (avrxmega6): Increase max flash segments from 5 to 6.
15555 * config/avr/t-multilib: Regenerate.
15556 * config/avr/avr-tables.opt: Regenerate.
15557 * doc/avr-mmcu.texi: Regenerate.
15559 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15561 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15562 (avr_device_to_arch): Rename to avr_device_to_ld.
15563 (avr_device_to_as): New prototype.
15564 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15565 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15566 * config/avr/driver-avr.c (avr_device_to_as): New.
15567 (avr_device_to_arch): Rename to avr_device_to_ld.
15569 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15571 PR middle-end/56461
15572 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15573 method on dr_chain and result_chain.
15575 PR middle-end/56461
15576 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15577 pointer_set_destroy on not_executed_last_iteration.
15579 PR middle-end/56461
15580 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15582 PR middle-end/56461
15583 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15584 FOR_EACH_DEFINED_FUNCTION when freeing state.
15586 PR middle-end/56461
15587 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15589 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15592 PR middle-end/56461
15593 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15594 known_aggs[i].items and release known_aggs vector.
15596 PR middle-end/56461
15597 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15599 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15601 * config/microblaze/microblaze.c (microblaze_emit_compare):
15602 Use xor for EQ/NE comparisions.
15603 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15604 (cbranchsf4): Adjust operator to comparison_operator.
15606 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15608 PR middle-end/56461
15609 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15611 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15612 vec_safe_push, always update *slot.
15613 (redirect_edge_var_map_clear): Use vec_free.
15614 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15615 (free_var_map_entry): Use vec_free.
15616 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15617 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15619 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15621 PR middle-end/45472
15622 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15623 when the may_trap_p bit of the exprs being merged differs.
15624 Reorder tests for speculativeness in the logical and operator.
15626 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15628 * incpath.c (add_standard_paths): Use reconcat instead of concat
15629 where appropriate and avoid leaking memory.
15631 * opts.h: Include obstack.h.
15632 (opts_concat): New prototype.
15633 (opts_obstack): New declaration.
15634 * opts.c (opts_concat): New function.
15635 (opts_obstack): New variable.
15636 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15637 (finish_options): Use opts_concat instead of concat
15638 and XOBNEWVEC instead of XNEWVEC.
15639 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15640 generate_option): Likewise.
15641 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15642 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15645 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15646 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15648 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15650 PR middle-end/56461
15651 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15653 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15655 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15656 (arm_block_move_unaligned_straight): Likewise.
15657 (arm_adjust_block_mem): Likewise.
15659 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15662 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15663 temp, cond and label.
15664 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15667 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15668 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15671 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15672 of prologue/epilogue functions.
15675 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15676 Remove unused variable cfa_offset.
15677 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15680 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15681 type promotion to unsigned.
15684 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15685 for HOST_WIDE_INT of 32 bit / same size as int.
15686 (arm_block_move_unaligned_straight): Likewise.
15687 (arm_adjust_block_mem): Likewise.
15690 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15693 2013-02-26 Marek Polacek <polacek@redhat.com>
15695 PR tree-optimization/56426
15696 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15698 2013-02-26 Richard Biener <rguenther@suse.de>
15701 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15702 unused variable loops.
15704 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15706 PR tree-optimization/56448
15707 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15708 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15709 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15710 later operands of the references, or even first operand for
15711 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15713 PR tree-optimization/56443
15714 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15715 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15716 to type_for_mode langhook.
15718 2013-02-25 Matt Turner <mattst88@gmail.com>
15720 * doc/invoke.texi: Document r4700.
15722 2013-02-25 Richard Biener <rguenther@suse.de>
15724 PR tree-optimization/56175
15725 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15727 (simplify_bitwise_binary): ... here. Also guard the conversion
15728 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15730 2013-02-25 Catherine Moore <clm@codesourcery.com>
15733 2013-02-24 Catherine Moore <clm@codesourcery.com>
15734 Maciej W. Rozycki <macro@codesourcery.com>
15735 Tom de Vries <tom@codesourcery.com>
15736 Nathan Sidwell <nathan@codesourcery.com>
15737 Iain Sandoe <iain@codesourcery.com>
15738 Nathan Froyd <froydnj@codesourcery.com>
15739 Chao-ying Fu <fu@mips.com>
15741 * doc/extend.texi: (micromips, nomicromips, nocompression):
15742 Document new function attributes.
15743 * doc/invoke.texi (minterlink-compressed, mmicromips,
15744 m14k, m14ke, m14kec): Document new options.
15745 (minterlink-mips16): Update documentation.
15746 * doc/md.texi (ZC, ZD): Document new constraints.
15747 * configure.ac (gcc_cv_as_micromips): Check if linker
15748 supports the .set micromips directive.
15749 * configure: Regenerate.
15750 * config.in: Regenerate.
15751 * config/mips/mips-tables.opt: Regenerate.
15752 * config/mips/micromips.md: New file.
15753 * constraints.md (ZC, AD): New constraints.
15754 * config/mips/predicates.md (movep_src_register): New predicate.
15755 (movep_src_operand): New predicate.
15756 (non_volatile_mem_operand): New predicate.
15757 * config/mips/mips.md (multimem): New type.
15758 (length): Differentiate between 17-bit and 18-bit branch offsets.
15759 (MOVEP1, MOVEP2): New mode iterator.
15760 (mov_<load>l): Use ZC constraint.
15761 (mov_<load>r): Likewise.
15762 (mov_<store>l): Likewise.
15763 (mov_<store>r): Likewise.
15764 (*branch_equality<mode>_inverted): Add microMIPS support.
15765 (*branch_equality<mode>): Likewise.
15766 (*jump_absolute): Likewise.
15767 (indirect_jump_<mode>): Likewise.
15768 (tablejump_<mode>): Likewise.
15769 (<optab>_internal): Likewise.
15770 (sibcall_internal): Likewise.
15771 (sibcall_value_internal): Likewise.
15772 (prefetch): Use constraint ZD.
15773 * config/mips/mips.opt (minterlink-compressed): New option.
15774 (minterlink-mips16): Now an alias for minterlink-compressed.
15775 (mmicromips): New option.
15776 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15777 (compare_and_swap_12): Likewise.
15778 (sync_add<mode>): Likewise.
15779 (sync_<optab>_12): Likewise.
15780 (sync_old_<optab>_12): Likewise.
15781 (sync_new_<optab>_12): Likewise.
15782 (sync_nand_12): Likewise.
15783 (sync_old_nand_12): Likewise.
15784 (sync_new_nand_12): Likewise.
15785 (sync_sub<mode>): Likewise.
15786 (sync_old_add<mode>): Likewise.
15787 (sync_old_sub<mode>): Likewise.
15788 (sync_new_add<mode>): Likewise.
15789 (sync_new_sub<mode>): Likewise.
15790 (sync_<optab><mode>): Likewise.
15791 (sync_old_<optab><mode>): Likewise.
15792 (sync_new_<optab><mode>): Likewise.
15793 (sync_nand<mode>): Likewise.
15794 (sync_old_nand<mode>): Likewise.
15795 (sync_new_nand<mode>): Likewise.
15796 (sync_lock_test_and_set<mode>): Likewise.
15797 (test_and_set_12): Likewise.
15798 (atomic_compare_and_swap<mode>): Likewise.
15799 (atomic_exchange<mode>_llsc): Likewise.
15800 (atomic_fetch_add<mode>_llsc): Likewise.
15801 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15802 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15803 (umips_save_restore_pattern_p): Likewise.
15804 (umips_load_store_pair_p): Likewise.
15805 (umips_output_load_store_pair): Likewise.
15806 (umips_movep_target_p): Likewise.
15807 (umips_12bit_offset_address_p): Likewise.
15808 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15809 (mips_base_mips16): Rename this...
15810 (mips_base_compression_flags): ...to this. Update all uses.
15811 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15812 (mips_mips16_decl_p): Delete.
15813 (mips_nomips16_decl_p): Delete.
15814 (mips_get_compress_on_flags): New function.
15815 (mips_get_compress_off_flags): New function.
15816 (mips_get_compress_mode): New function.
15817 (mips_get_compress_on_name): New function.
15818 (mips_get_compress_off_name): New function.
15819 (mips_insert_attributes): Support multiple compression types.
15820 (mips_merge_decl_attributes): Likewise.
15821 (umips_12bit_offset_address_p): New function.
15822 (mips_start_function_definition): Emit .set micromips directive.
15823 (mips_call_may_need_jalx_p): New function.
15824 (mips_function_ok_for_sibcall): Add microMIPS support.
15825 (mips_print_operand_punctuation): Support short delay slots and
15827 (umips_swm_mask, umips_swm_encoding): New.
15828 (umips_build_save_restore): New function.
15829 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15830 (was_mips16_p): Remove.
15831 (old_compression_mode): New.
15832 (mips_set_compression_mode): New function.
15833 (mips_set_current_function): Add microMIPS support.
15834 (mips_option_override): Likewise.
15835 (umips_save_restore_pattern_p): New function.
15836 (umips_output_save_restore): New function.
15837 (umips_load_store_pair_p_1): New function.
15838 (umips_load_store_pair_p): New function.
15839 (umips_output_load_store_pair_1): New function.
15840 (umips_output_load_store_pair): New function.
15841 (umips_movep_target_p) New function.
15842 (mips_prepare_pch_save): Add microMIPS support.
15843 * config/mips/mips.h (TARGET_COMPRESSION): New.
15844 (TARGET_CPU_CPP_BUILTINS): Update macro
15845 to use new compression flags and to support microMIPS.
15846 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15847 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15848 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15849 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15850 (ASM_SPEC): Support mmicromips and mno-micromips.
15851 (M16STORE_REG_P): New macro.
15852 (MIPS_CALL): Support TARGET_MICROMIPS.
15853 (MICROMIPS_J): New macro.
15854 (mips_base_mips16): Rename this...
15855 (mips_base_compression_flags): ...to this.
15856 (UMIPS_12BIT_OFFSET_P): New macro.
15857 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15858 (MULTILIB_DIRNAMES): Likewise.
15860 2013-02-25 Tom de Vries <tom@codesourcery.com>
15862 PR rtl-optimization/56131
15863 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15864 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15865 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15867 2013-02-25 Tobias Burnus <burnus@net-b.de>
15869 * doc/invoke.texi (-fsanitize=): Move from optimization
15870 to debugging options.
15872 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15874 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15876 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15877 Alexander Monakov <amonakov@ispras.ru>
15879 PR middle-end/56077
15880 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15881 flush pending lists also on non-jumps. Adjust comment.
15883 2013-02-24 Catherine Moore <clm@codesourcery.com>
15884 Maciej W. Rozycki <macro@codesourcery.com>
15885 Tom de Vries <tom@codesourcery.com>
15886 Nathan Sidwell <nathan@codesourcery.com>
15887 Iain Sandoe <iain@codesourcery.com>
15888 Nathan Froyd <froydnj@codesourcery.com>
15889 Chao-ying Fu <fu@mips.com>
15891 * doc/extend.texi: (micromips, nomicromips, nocompression):
15892 Document new function attributes.
15893 * doc/invoke.texi (minterlink-compressed, mmicromips,
15894 m14k, m14ke, m14kec): Document new options.
15895 (minterlink-mips16): Update documentation.
15896 * doc/md.texi (ZC, ZD): Document new constraints.
15897 * configure.ac (gcc_cv_as_micromips): Check if linker
15898 supports the .set micromips directive.
15899 * configure: Regenerate.
15900 * config.in: Regenerate.
15901 * config/mips/mips-tables.opt: Regenerate.
15902 * config/mips/micromips.md: New file.
15903 * constraints.md (ZC, AD): New constraints.
15904 * config/mips/predicates.md (movep_src_register): New predicate.
15905 (movep_src_operand): New predicate.
15906 (non_volatile_mem_operand): New predicate.
15907 * config/mips/mips.md (multimem): New type.
15908 (length): Differentiate between 17-bit and 18-bit branch offsets.
15909 (MOVEP1, MOVEP2): New mode iterator.
15910 (mov_<load>l): Use ZC constraint.
15911 (mov_<load>r): Likewise.
15912 (mov_<store>l): Likewise.
15913 (mov_<store>r): Likewise.
15914 (*branch_equality<mode>_inverted): Add microMIPS support.
15915 (*branch_equality<mode>): Likewise.
15916 (*jump_absolute): Likewise.
15917 (indirect_jump_<mode>): Likewise.
15918 (tablejump_<mode>): Likewise.
15919 (<optab>_internal): Likewise.
15920 (sibcall_internal): Likewise.
15921 (sibcall_value_internal): Likewise.
15922 (prefetch): Use constraint ZD.
15923 * config/mips/mips.opt (minterlink-compressed): New option.
15924 (minterlink-mips16): Now an alias for minterlink-compressed.
15925 (mmicromips): New option.
15926 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15927 (compare_and_swap_12): Likewise.
15928 (sync_add<mode>): Likewise.
15929 (sync_<optab>_12): Likewise.
15930 (sync_old_<optab>_12): Likewise.
15931 (sync_new_<optab>_12): Likewise.
15932 (sync_nand_12): Likewise.
15933 (sync_old_nand_12): Likewise.
15934 (sync_new_nand_12): Likewise.
15935 (sync_sub<mode>): Likewise.
15936 (sync_old_add<mode>): Likewise.
15937 (sync_old_sub<mode>): Likewise.
15938 (sync_new_add<mode>): Likewise.
15939 (sync_new_sub<mode>): Likewise.
15940 (sync_<optab><mode>): Likewise.
15941 (sync_old_<optab><mode>): Likewise.
15942 (sync_new_<optab><mode>): Likewise.
15943 (sync_nand<mode>): Likewise.
15944 (sync_old_nand<mode>): Likewise.
15945 (sync_new_nand<mode>): Likewise.
15946 (sync_lock_test_and_set<mode>): Likewise.
15947 (test_and_set_12): Likewise.
15948 (atomic_compare_and_swap<mode>): Likewise.
15949 (atomic_exchange<mode>_llsc): Likewise.
15950 (atomic_fetch_add<mode>_llsc): Likewise.
15951 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15952 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15953 (umips_save_restore_pattern_p): Likewise.
15954 (umips_load_store_pair_p): Likewise.
15955 (umips_output_load_store_pair): Likewise.
15956 (umips_movep_target_p): Likewise.
15957 (umips_12bit_offset_address_p): Likewise.
15958 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15959 (mips_base_mips16): Rename this...
15960 (mips_base_compression_flags): ...to this. Update all uses.
15961 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15962 (mips_mips16_decl_p): Delete.
15963 (mips_nomips16_decl_p): Delete.
15964 (mips_get_compress_on_flags): New function.
15965 (mips_get_compress_off_flags): New function.
15966 (mips_get_compress_mode): New function.
15967 (mips_get_compress_on_name): New function.
15968 (mips_get_compress_off_name): New function.
15969 (mips_insert_attributes): Support multiple compression types.
15970 (mips_merge_decl_attributes): Likewise.
15971 (umips_12bit_offset_address_p): New function.
15972 (mips_start_function_definition): Emit .set micromips directive.
15973 (mips_call_may_need_jalx_p): New function.
15974 (mips_function_ok_for_sibcall): Add microMIPS support.
15975 (mips_print_operand_punctuation): Support short delay slots and
15977 (umips_swm_mask, umips_swm_encoding): New.
15978 (umips_build_save_restore): New function.
15979 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15980 (was_mips16_p): Remove.
15981 (old_compression_mode): New.
15982 (mips_set_compression_mode): New function.
15983 (mips_set_current_function): Add microMIPS support.
15984 (mips_option_override): Likewise.
15985 (umips_save_restore_pattern_p): New function.
15986 (umips_output_save_restore): New function.
15987 (umips_load_store_pair_p_1): New function.
15988 (umips_load_store_pair_p): New function.
15989 (umips_output_load_store_pair_1): New function.
15990 (umips_output_load_store_pair): New function.
15991 (umips_movep_target_p) New function.
15992 (mips_prepare_pch_save): Add microMIPS support.
15993 * config/mips/mips.h (TARGET_COMPRESSION): New.
15994 (TARGET_CPU_CPP_BUILTINS): Update macro
15995 to use new compression flags and to support microMIPS.
15996 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15997 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15998 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15999 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16000 (ASM_SPEC): Support mmicromips and mno-micromips.
16001 (M16STORE_REG_P): New macro.
16002 (MIPS_CALL): Support TARGET_MICROMIPS.
16003 (MICROMIPS_J): New macro.
16004 (mips_base_mips16): Rename this...
16005 (mips_base_compression_flags): ...to this.
16006 (UMIPS_12BIT_OFFSET_P): New macro.
16007 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16008 (MULTILIB_DIRNAMES): Likewise.
16010 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16013 * target-globals.c (save_target_globals): For init_reg_sets and
16014 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16016 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16018 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16019 * config/aarch64/t-aarch64
16020 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16022 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16024 PR inline-asm/56148
16025 * lra-constraints.c (process_alt_operands): Reload operand
16026 conflicting with earlier clobber only if no more other conflicting
16029 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16032 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16033 if not linking a shared library.
16035 2013-02-22 Seth LaForge <sethml@google.com>
16037 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16039 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16041 * config/arm/arm.md (split for extendsidi): Update condition.
16042 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16043 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16044 (qhs_zextenddi_cstr): Likewise.
16046 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16048 PR middle-end/56420
16049 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16050 avoid signed wrapping.
16051 (expand_mult): Handle properly multiplication by
16052 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16053 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16054 in the compiler if coeff is HOST_WIDE_INT_MIN.
16055 (expand_divmod): Don't make ext_op1 static, change it's type to
16056 uhwi. Avoid undefined behavior in -INTVAL (op1).
16058 PR rtl-optimization/50339
16059 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16061 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16062 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16063 into splitting_ashiftrt field.
16064 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16066 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16069 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16071 PR middle-end/56108
16072 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16073 are sure to go irrevocable.
16075 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16077 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16078 scalars are valid operands.
16080 2013-02-21 Martin Jambor <mjambor@suse.cz>
16082 PR tree-optimization/56310
16083 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16084 only matching indices and non-negative final offsets.
16085 (intersect_aggregates_with_edge): Pass src_idx to
16086 agg_replacements_to_vector. Pass src_idx insstead of index to
16087 intersect_with_agg_replacements.
16089 2013-02-21 Martin Jambor <mjambor@suse.cz>
16091 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16092 instead of hard-wired defaults.
16094 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16096 * doc/invoke.texi (MIPS Options): Update documentation of the
16097 floating-point multiply-accumulate instruction restrictions.
16099 2013-02-21 Kostya Serebryany <kcc@google.com>
16101 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16102 asan_shadow_offset on x86_64 linux.
16104 2013-02-21 Richard Biener <rguenther@suse.de>
16106 PR tree-optimization/56415
16108 2013-02-11 Richard Biener <rguenther@suse.de>
16110 PR tree-optimization/56273
16111 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16114 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16117 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16120 PR inline-asm/56405
16121 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16122 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16124 2013-02-20 Jan Hubicka <jh@suse.cz>
16126 PR tree-optimization/56265
16127 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16128 when target is referenced for first time.
16130 2013-02-20 Richard Biener <rguenther@suse.de>
16132 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16133 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16134 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16135 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16136 not return anything.
16137 (rename_ssa_copies): Do not remove unused locals.
16138 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16139 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16140 * passes.c (execute_function_todo): Do not schedule unused locals
16141 removal if cleanup_tree_cfg did something.
16142 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16143 about the number of removed locals.
16145 2013-02-20 Richard Biener <rguenther@suse.de>
16147 PR tree-optimization/56398
16148 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16150 2013-02-20 Martin Jambor <mjambor@suse.cz>
16152 PR tree-optimization/55334
16153 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16154 restricted pointers to arrays.
16156 2013-02-20 Richard Biener <rguenther@suse.de>
16157 Jakub Jelinek <jakub@redhat.com>
16159 PR tree-optimization/56396
16160 * tree-ssa-ccp.c (n_const_val): New static variable.
16161 (get_value): Return NULL for SSA names we don't have a lattice
16163 (ccp_initialize): Initialize n_const_val.
16164 * tree-ssa-copy.c (n_copy_of): New static variable.
16165 (init_copy_prop): Initialize n_copy_of.
16166 (get_value): Return NULL_TREE for SSA names we don't have a
16169 2013-02-20 Martin Jambor <mjambor@suse.cz>
16171 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16173 2013-02-20 Richard Biener <rguenther@suse.de>
16175 * genpreds.c (write_lookup_constraint): Do not compare first
16176 letter of the constraint again.
16178 2013-02-20 Richard Biener <rguenther@suse.de>
16180 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16182 (get_use_iv_cost): Terminate hashtable walk when coming across
16185 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16187 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16188 reassociation for avx2 targets.
16190 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16192 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16193 Add version check for v8.10.a to enable microblaze_has_clz
16194 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16195 version and TARGET_PATTERN_COMPARE check
16196 * config/microblaze/microblaze.md: New clzsi2 instruction
16198 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16200 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
16201 function before branching.
16203 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16205 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
16207 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
16209 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16211 PR middle-end/55889
16212 * sel-sched.c: Include ira.h.
16213 (implicit_clobber_conflict_p): New function.
16214 (moveup_expr): Use it.
16215 * Makefile.in (sel-sched.o): Depend on ira.h.
16217 2013-02-19 Richard Biener <rguenther@suse.de>
16219 PR tree-optimization/56384
16220 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
16221 (vn_hash_type): Split out from ...
16222 (vn_hash_constant_with_type): ... here.
16223 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
16224 (vn_phi_eq): Compare types from vn_phi_s structure.
16225 (vn_phi_lookup): Populate vn_phi_s type.
16226 (vn_phi_insert): Likewise.
16228 2013-02-19 Jakub Jelinek <jakub@redhat.com>
16230 PR tree-optimization/56350
16231 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
16232 if haven't found reduction or nested cycle operand, rather than
16233 asserting we must find it.
16235 PR tree-optimization/56381
16236 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
16239 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
16240 Jakub Jelinek <jakub@redhat.com>
16243 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
16244 (swap_optab_enable): Same.
16245 (init_all_optabs): Use argument instead of global.
16246 * tree.h (struct tree_optimization_option): New field target_optabs.
16247 * expr.h (init_all_optabs): Add argument to prototype.
16248 (TREE_OPTIMIZATION_OPTABS): New.
16249 (save_optabs_if_changed): Protoize.
16250 * optabs.h: Declare this_fn_optabs.
16251 * optabs.c (save_optabs_if_changed): New.
16252 Declare this_fn_optabs.
16253 (init_optabs): Add argument to init_all_optabs() call.
16254 * function.c (invoke_set_current_function_hook): Handle per
16256 * function.h (struct function): New field optabs.
16257 * config/mips/mips.c (mips_set_mips16_mode): Handle when
16258 optimization_current_node has changed.
16259 * target-globals.h (save_target_globals_default_opts): Protoize.
16260 * target-globals.c (save_target_globals_default_opts): New.
16262 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16265 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
16266 registers %fr12 and %fr12R as call used.
16269 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
16270 and HImode, require all displacements to be an integer multiple of
16272 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
16273 only allow QImode and HImode when reload is in progress and strict is
16274 true. Likewise for symbolic addresses. Use base14_operand to check
16275 displacements in REG+BASE addresses.
16277 2013-02-18 Richard Biener <rguenther@suse.de>
16279 PR tree-optimization/56366
16280 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16281 handle sign-conversion of outer-loop initial induction value.
16283 2013-02-18 Richard Biener <rguenther@suse.de>
16285 PR middle-end/56349
16286 * cfghooks.c (merge_blocks): If we merge a latch into another
16287 block adjust references to it.
16288 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
16289 (verify_loop_structure): Verify that a recorded latch is in fact
16292 2013-02-18 Richard Biener <rguenther@suse.de>
16294 PR tree-optimization/56321
16295 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
16296 order SSA name release and virtual operand unlinking.
16298 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16300 * config/microblaze/microblaze.md (save_stack_block): Define.
16301 (restore_stack_block): Likewise.
16303 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16305 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
16306 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
16307 * config/microblaze/microblaze.c (microblaze_option_override):
16308 Bail out early for PIC modes when target does not support PIC.
16310 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16312 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
16313 Replace with a microblaze version.
16314 (microblaze_trampoline_init): Adapt for microblaze.
16315 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
16318 2013-02-16 Jakub Jelinek <jakub@redhat.com>
16319 Dodji Seketeli <dodji@redhat.com>
16322 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
16323 (instrument_mem_region_access): Do not forget to always put
16324 instrumentation of the of 'base' and 'base + len' in a "if (len !=
16325 0) statement, even for cases where either 'base' or 'base + len'
16326 are not instrumented -- because they have been previously
16327 instrumented. Simplify the logic by putting all the statements
16328 instrument 'base + len' inside a sequence, and then insert that
16329 sequence right before the current insertion point. Then, to
16330 instrument 'base + len', just get an iterator on that statement.
16331 And do not forget to update the pointer to iterator the function
16332 received as argument.
16334 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
16336 PR rtl-optimization/56348
16337 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
16339 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
16341 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
16342 (clean_graph_dump_file): Pass base to start_graph_dump.
16344 2013-02-14 Richard Henderson <rth@redhat.com>
16347 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
16349 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
16351 * collect2-aix.h: Define F_LOADONLY.
16353 2013-02-14 Richard Biener <rguenther@suse.de>
16356 * varasm.c (output_constant_def_1): Get the decl representing
16357 the constant as argument.
16358 (output_constant_def): Wrap output_constant_def_1.
16359 (make_decl_rtl): Use output_constant_def_1 with the decl
16360 representing the constant.
16361 (build_constant_desc): Optionally re-use a decl already
16362 representing the constant.
16363 (tree_output_constant_def): Adjust.
16365 2013-02-14 Dodji Seketeli <dodji@redhat.com>
16368 * asan.c (instrument_builtin_call): Really put the length of the
16369 second source argument into src1_len.
16371 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16373 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
16374 argument. If it is false, don't create edge from then_bb to
16376 (insert_if_then_before_iter): Pass true to it.
16377 (build_check_stmt): Pass false to it.
16378 (transform_statements): Flush hash table only on extended basic
16379 block boundaries, rather than at the beginning of every bb.
16380 Don't flush hash table on nonfreeing_call_p calls.
16381 * tree-flow.h (nonfreeing_call_p): New prototype.
16382 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
16384 2013-02-13 David S. Miller <davem@davemloft.net>
16386 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
16388 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
16391 * ira.c (max_regno_before_ira): Move from ...
16393 (fix_reg_equiv_init): Use max_regno_before_ira instead of
16396 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16398 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
16400 2013-02-13 Richard Biener <rguenther@suse.de>
16403 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
16404 globals in MEM_REFs.
16406 2013-02-13 Richard Biener <rguenther@suse.de>
16408 * loop-init.c (loop_optimizer_init): Clear loop state when
16409 re-initializing preserved loops.
16410 * loop-unswitch.c (unswitch_single_loop): Return whether
16411 we unswitched the loop. Do not verify loop state here.
16412 (unswitch_loops): When we unswitched a loop discover new loops.
16414 2013-02-13 Kostya Serebryany <kcc@google.com>
16416 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
16418 * sanitizer.def: Rename __asan_init to __asan_init_v1.
16420 2013-02-12 Dodji Seketeli <dodji@redhat.com>
16422 Avoid instrumenting duplicated memory access in the same basic block
16423 * Makefile.in (asan.o): Add new dependency on hash-table.h
16424 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
16425 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
16426 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
16427 (free_mem_ref_resources, has_mem_ref_been_instrumented)
16428 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
16429 (get_mem_ref_of_assignment): New functions.
16430 (get_mem_refs_of_builtin_call): Extract from
16431 instrument_builtin_call and tweak a little bit to make it fit with
16433 (instrument_builtin_call): Use the new
16434 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
16435 of is_gimple_builtin_call.
16436 (instrument_derefs, instrument_mem_region_access): Insert the
16437 instrumented memory reference into the hash table.
16438 (maybe_instrument_assignment): Renamed instrument_assignment into
16439 this, and change it to advance the iterator when instrumentation
16440 actually happened and return true in that case. This makes it
16441 homogeneous with maybe_instrument_assignment, and thus give a
16442 chance to callers to be more 'regular'.
16443 (transform_statements): Clear the memory reference hash table
16444 whenever we enter a new BB, when we cross a function call, or when
16445 we are done transforming statements. Use
16446 maybe_instrument_assignment instead of instrumentation. No more
16447 need to special case maybe_instrument_assignment and advance the
16448 iterator after calling it; it's now handled just like
16449 maybe_instrument_call. Update comment.
16451 2013-02-13 Richard Biener <rguenther@suse.de>
16453 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
16454 Fix loop discovery code.
16456 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
16458 PR inline-asm/56148
16459 * lra-constraints.c (process_alt_operands): Match early clobber
16460 operand with itself. Check conflicts with earlyclobber only if
16461 the operand is not reloaded. Prefer to reload conflicting operand
16462 if earlyclobber and matching operands are the same.
16464 2013-02-12 Richard Biener <rguenther@suse.de>
16467 * lto-streamer-out.c (write_symbol): Do not output symbols
16468 for hard register variables.
16470 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16473 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16474 (umulsidi3_insn, mulsidi3_insn): New insns.
16476 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16478 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16479 (struct tune_params): Add vec_costs field.
16480 * config/arm/arm.c (arm_builtin_vectorization_cost)
16481 (arm_add_stmt_cost): New functions.
16482 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16483 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16484 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16485 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16486 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16487 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16488 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16490 2013-02-12 Richard Biener <rguenther@suse.de>
16493 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16494 decls again if possible.
16496 2013-02-12 Richard Biener <rguenther@suse.de>
16498 PR middle-end/56288
16499 * tree-ssa.c (verify_ssa_name): Fix check, move
16500 SSA_NAME_IN_FREE_LIST check up.
16502 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16503 Steven Bosscher <steven@gcc.gnu.org>
16505 PR rtl-optimization/56151
16506 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16507 equal to op0 or op1, and last_insn pattern is CODE operation
16508 with MEM dest and one of the operands matches that MEM.
16510 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16512 * doc/extend.texi: Document Function Multiversioning and "default"
16513 parameter string to target attribute.
16514 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16515 target attribute parameter is "default".
16516 (ix86_compare_version_priority): Remove checks for target attribute.
16517 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16518 Remove check for target attribute equal to NULL. Add assert.
16519 (ix86_generate_version_dispatcher_body): Change error to sorry.
16521 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16522 Jack Howarth <howarth@bromo.med.uc.edu>
16523 Patrick Marlier <patrick.marlier@gmail.com>
16526 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16527 define ENDFILE_SPEC as TM_DESTRUCTOR.
16528 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16530 2013-02-11 Alexander Potapenko <glider@google.com>
16531 Jack Howarth <howarth@bromo.med.uc.edu>
16532 Jakub Jelinek <jakub@redhat.com>
16535 * config/darwin.c (cdtor_record): Rename ctor_record.
16536 (sort_cdtor_records): Rename sort_ctor_records.
16537 (finalize_dtors): New routine to sort destructors by
16538 priority before use in assemble_integer.
16539 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16541 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16543 PR rtl-optimization/56275
16544 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16545 offset is non-negative and less than cmode size before
16546 calling simplify_subreg.
16548 2013-02-11 Richard Biener <rguenther@suse.de>
16550 PR tree-optimization/56264
16551 * cfgloop.h (fix_loop_structure): Adjust prototype.
16552 * loop-init.c (fix_loop_structure): Return the number of
16553 newly discovered loops.
16554 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16555 are discovered, do a full loop-closed SSA rewrite.
16557 2013-02-11 Richard Biener <rguenther@suse.de>
16559 PR tree-optimization/56273
16560 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16562 (check_array_ref): Fix missing newline in dumps.
16563 (search_for_addr_array): Likewise.
16565 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16567 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16569 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16572 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16574 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16576 PR rtl-optimization/56246
16577 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16579 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16580 constraints are satisfied.
16582 2013-02-08 Jeff Law <law@redhat.com>
16585 * emit-rtl.c (reg_is_parm_p): New function.
16586 * regs.h (reg_is_parm_p): New prototype.
16587 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16588 callee-clobbered registers.
16590 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16593 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16594 If there is no implicit builtin declaration, just return NULL.
16596 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16598 * config/i386/sse.md (FMAMODEM): New mode iterator.
16599 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16600 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16602 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16604 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16605 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16606 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16608 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16610 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16611 (microblaze*-*-elf): Likewise.
16612 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16614 * config/microblaze/microblaze-c.c: Add builtin defines for
16615 _LITTLE_ENDIAN and _BIG_ENDIAN.
16616 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16617 add to TARGET_DEFAULT flags.
16618 Expand ASM_SPEC and LINK_SPEC.
16619 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16620 * config/microblaze/microblaze.md: Update extendsidi2 and
16621 movdi_internal instructions to use low-order / high-order reg
16623 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16624 options and inversemask / mask of LITTLE_ENDIAN.
16625 * config/microblaze/t-microblaze: Expand multilib options to
16626 include mlittle-endian (le) and update exceptions patterns.
16628 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16630 PR rtl-optimization/56195
16631 * lra-constraints.c (get_reload_reg): Don't reuse regs
16632 if they have smaller mode than requested, if they have
16633 wider mode than requested, try to return a SUBREG.
16635 PR tree-optimization/56250
16636 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16637 if type is unsigned and code isn't MULT_EXPR.
16639 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16641 PR tree-optimization/56064
16642 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16643 bits according to mode.
16644 * fixed-value.h (fixed_from_double_int)
16645 (const_fixed_from_double_int): Adjust comments.
16647 2013-02-08 Richard Biener <rguenther@suse.de>
16650 * lto-streamer.h (struct data_in): Remove current_file, current_line
16651 and current_col members.
16652 * lto-streamer-out.c (lto_output_location): Stream changed bits
16653 en-block for efficiency.
16654 * lto-streamer-in.c (clear_line_info): Remove.
16655 (lto_input_location): Cache current file, line and column
16656 globally via local statics. Read changed bits en-block.
16657 (input_function): Do not call clear_line_info.
16658 (lto_read_body): Likewise.
16659 (lto_input_toplevel_asms): Likewise.
16661 2013-02-08 Michael Matz <matz@suse.de>
16663 PR tree-optimization/52448
16664 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16665 (nt_call_phase): New static.
16666 (add_or_mark_expr): Only mark accesses with newer phase than any
16668 (nonfreeing_call_p): New.
16669 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16670 (nt_fini_block): Keep blocks marked as visited.
16671 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16673 2013-02-08 Richard Biener <rguenther@suse.de>
16675 * ira.c (ira): Free broken dominator information.
16677 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16679 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16681 2013-02-08 Marek Polacek <polacek@redhat.com>
16683 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16685 2013-02-08 Richard Biener <rguenther@suse.de>
16687 PR middle-end/56181
16688 * cfgloop.h (flow_loops_find): Adjust.
16689 (bb_loop_header_p): Declare.
16690 * cfgloop.c (bb_loop_header_p): New function split out from ...
16691 (flow_loops_find): ... here. Adjust function signature,
16692 support incremental loop structure update.
16693 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16694 * cfgloopmanip.c (fix_loop_structure): Move ...
16695 * loop-init.c (fix_loop_structure): ... here.
16696 (apply_loop_flags): Split out from ...
16697 (loop_optimizer_init): ... here.
16698 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16699 in incremental mode, only remove dead loops here.
16701 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16704 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16705 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16706 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16707 (*round<mode>3.libgcc): New insns for fixed-modes.
16708 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16709 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16710 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16711 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16712 implementations. Define to __builtin_avr_absFX,
16713 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16714 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16715 __builtin_avr_countlsFX, respectively.
16716 * config/avr/avr-c.c (target.h): Include it.
16717 (enum avr_builtin_id): New enum.
16718 (avr_resolve_overloaded_builtin): New static function.
16719 (avr_register_target_pragmas): Use it to set
16720 targetm.resolve_overloaded_builtin.
16721 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16722 tree nodes used by DEF_BUILTIN.
16723 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16724 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16725 <AVR_BUILTIN_xxBITS>: Same.
16727 2013-02-08 Richard Biener <rguenther@suse.de>
16729 * cfgloop.c (verify_loop_structure): Properly handle
16730 a loop exiting to another loop header.
16731 * ira-int.h (ira_loops): Remove.
16732 * ira.c (ira_loops): Remove.
16733 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16734 (do_reload): Use loop_optimizer_finalize.
16735 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16736 number_of_loops to access the loop tree.
16737 (more_one_region_p): Likewise.
16738 (finish_loop_tree_nodes): Likewise.
16739 (rebuild_regno_allocno_maps): Likewise.
16740 (mark_loops_for_removal): Likewise.
16741 (mark_all_loops_for_removal): Likewise.
16742 (remove_unnecessary_regions): Likewise.
16743 (ira_build): Likewise.
16744 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16746 2013-02-08 Richard Biener <rguenther@suse.de>
16748 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16749 * ipa-pure-const.c (analyze_function): Avoid calling
16750 mark_irreducible_loops twice.
16751 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16753 2013-02-07 David S. Miller <davem@davemloft.net>
16755 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16757 * var-tracking.c (vt_add_function_parameter): Test the presence of
16758 HAVE_window_save properly and do not remap argument registers when
16759 we have a leaf function.
16761 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16764 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16766 * config/i386/i386.c (ix86_print_operand): Ditto.
16768 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16770 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16772 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16774 PR rtl-optimization/56225
16775 * lra-constraints.c (process_alt_operands): Check that reload hard
16776 reg can hold value for strict_low_part.
16778 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16781 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16782 dwarf2out_end_function.
16783 (in_first_function_p, maybe_at_text_label_p,
16784 first_loclabel_num_not_at_text_label): New variables.
16785 (dwarf2out_var_location): In the first function find out
16786 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16787 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16790 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16792 PR rtl-optimization/56178
16793 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16794 SUBREG of a register. Tidy up related block of code.
16795 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16796 note if the source is a register or a SUBREG of a register.
16798 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16801 * config/rs6000/rs6000.md (ptrm): New mode attr.
16802 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16803 call_value_indirect_aix<pttrsize>,
16804 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16807 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16809 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16810 if -bnortl. Convert to strcmp and strncmp.
16812 2013-02-07 Alan Modra <amodra@gmail.com>
16815 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16816 addresses won't wrap when offsetting.
16817 (rs6000_secondary_reload): Provide secondary reloads needed for
16818 wrapping LO_SUM addresses.
16820 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16822 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16823 MACH, just __MACH__.
16825 2013-02-06 Richard Biener <rguenther@suse.de>
16827 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16828 instead of calling fix_loop_structure.
16830 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16832 PR middle-end/56217
16833 * omp-low.c (use_pointer_for_field): Return false if
16834 lower_send_shared_vars doesn't generate any copy-out code.
16836 2013-02-06 Tom de Vries <tom@codesourcery.com>
16838 PR rtl-optimization/56131
16839 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16840 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16841 of the label is NULL. Add comment.
16843 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16845 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16848 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16849 (STATIC_LIBTSAN_LIBS): Likewise.
16850 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16851 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16852 is defined, don't add anything else beyond that.
16853 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16854 (LINK_COMMAND_SPEC): Use them.
16856 PR tree-optimization/56205
16857 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16858 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16859 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16861 2013-02-05 Richard Biener <rguenther@suse.de>
16863 PR tree-optimization/53342
16864 PR tree-optimization/53185
16865 * tree-vectorizer.h (vect_check_strided_load): Remove.
16866 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16867 not disallow peeling for vectorized strided loads.
16868 (vect_check_strided_load): Make static and simplify.
16869 (vect_analyze_data_refs): Adjust.
16870 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16871 correctly when vectorizing strided loads.
16873 2013-02-05 Richard Biener <rguenther@suse.de>
16875 * doc/install.texi: Refer to ISL, not PPL.
16877 2013-02-05 Jan Hubicka <jh@suse.cz>
16879 PR tree-optimization/55789
16880 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16882 2013-02-05 Jan Hubicka <jh@suse.cz>
16884 PR tree-optimization/55789
16885 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16886 the dead call anyway.
16888 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16891 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16893 2013-02-04 Alexander Potapenko <glider@google.com>
16894 Jack Howarth <howarth@bromo.med.uc.edu>
16895 Jakub Jelinek <jakub@redhat.com>
16898 * config/darwin.c (sort_ctor_records): Stabilized qsort
16899 on constructor priority by using original position.
16900 (finalize_ctors): New routine to sort constructors by
16901 priority before use in assemble_integer.
16902 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16904 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16907 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16908 about visibility on artificial decls.
16909 * config/sol2.c (solaris_assemble_visibility): Likewise.
16911 2013-02-04 Kai Tietz <ktietz@redhat.com>
16914 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16915 argument and improve checking of return-argument types for 16-byte
16917 (ix86_function_value_1): Add additional valtype argument on call
16918 of function_value_64.
16919 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16920 handling infunction_value_64 function.
16922 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16924 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16926 2013-02-04 Richard Biener <rguenther@suse.de>
16928 PR tree-optimization/56188
16929 * tree-ssa-structalias.c (label_visit): Consider case with
16930 initially non-empty points-to set.
16931 (perform_var_substitution): Dump node mapping and clean up.
16933 2013-02-04 Richard Guenther <rguenther@suse.de>
16936 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16937 node prevail as last resort.
16938 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16939 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16941 2013-02-04 Richard Biener <rguenther@suse.de>
16943 PR tree-optimization/56113
16944 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16946 (equiv_class_lookup_or_add): ... this.
16947 (label_visit): Adjust and fix error in previous patch.
16948 (perform_var_substitution): Adjust.
16950 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16952 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16953 * config/sh/sh4-300.md: Likewise.
16954 * config/sh/sh4a.md: Likewise.
16955 * config/sh/constraints.md: Likewise.
16956 * config/sh/sh.md: Likewise.
16957 * config/sh/netbsd-elf.h: Likewise.
16958 * config/sh/predicates.md: Likewise.
16959 * config/sh/sh-protos.h: Likewise.
16960 * config/sh/ushmedia.h: Likewise.
16961 * config/sh/linux.h: Likewise.
16962 * config/sh/sh.c: Likewise.
16963 * config/sh/superh.h: Likewise.
16964 * config/sh/elf.h: Likewise.
16965 * config/sh/sh4.md: Likewise.
16966 * config/sh/sh.h: Likewise.
16968 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16970 * config/pa/constraints.md: Adjust unused letters. Change "T"
16971 constraint to match_test floating_point_store_memory_operand().
16972 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16973 (base14_operand): New.
16974 (floating_point_store_memory_operand): New.
16975 (integer_store_memory_operand): Revise to use base14_operand and
16976 reg_plus_base_memory_operand.
16977 (move_dest_operand): Allow symbolic_memory_operands.
16978 (symbolic_memory_operand): Check for LO_SOM.
16979 (symbolic_operand): Change default case to break.
16980 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16981 CONST_DOUBLE values to be reloaded by putting them into memory when
16982 the destination is a floating point register.
16983 (movdf): Remove code to handle CONST_DOUBLE.
16985 (reload_indf_r1): New.
16986 (reload_insf_r1): New.
16987 Consistently use "Q" and "T" constraints with integer and floating
16988 point move instructions, respectively.
16989 (movdi): Remove FAIL.
16990 Change predicate for source operand unamed DImode move from
16991 general_operand to move_src_operand.
16992 (umulsidi3): Change predicate for destination operand to
16994 Likewise for similar unamed patterns.
16995 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16996 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16997 (hppa_legitimize_address): Simplify mask calculation.
16998 (pa_emit_move_sequence): Revised handling of secondary reloads from
16999 REG+D addresses for floating point loads and stores. Directly handle
17000 loading CONST0_RTX (mode) to a floating point register.
17001 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17002 to floating point registers. Don't restrict secondary reloads to
17003 floating point registers to integer modes. Revise some comments and
17005 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17006 (pa_legitimate_address_p): New.
17007 (pa_legitimize_reload_address): New.
17008 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17009 (STRICT_REG_OK_FOR_BASE_P): New.
17010 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17011 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17013 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17014 Andrew Dixie <andrewd@gentrack.com>
17016 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17019 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17021 * expmed.c (extract_bit_field_1): Pass the full width of the
17022 structure to get_best_reg_extraction_insn.
17024 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17027 * configure.ac (use_cxa_atexit): Add AIX.
17028 * configure: Regenerate.
17030 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17032 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17035 * final.c (need_profile_function): New variable.
17036 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17037 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17038 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17039 notes, targetm.asm_out.function_prologue doesn't emit anything,
17040 HAVE_prologue and profiler should be emitted before prologue,
17041 set need_profile_function instead of emitting it.
17042 (final_scan_insn): If need_profile_function, emit
17043 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17044 NOTE_INSN_FUNCTION_BEG note.
17046 2013-02-01 Richard Henderson <rth@redhat.com>
17048 * config/rs6000/rs6000.md (smulditi3): New.
17051 * config/alpha/alpha.md (umulditi3): New.
17053 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17055 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17056 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17058 2013-02-01 Richard Biener <rguenther@suse.de>
17060 PR tree-optimization/56113
17061 * tree-ssa-structalias.c (label_visit): Reduce work for
17062 single-predecessor nodes.
17064 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17066 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17067 range isn't testing for zero.
17069 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17071 PR middle-end/56113
17072 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17074 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17075 Nick Clifton <nickc@redhat.com>
17077 * config/v850/constraints.md (Q): Define as a memory constraint.
17078 * config/v850/predicates.md (label_ref_operand): New predicate.
17079 (e3v5_shift_operand): New predicate.
17080 (ior_operator): New predicate.
17081 * config/v850/t-v850: Add e3v5 multilib.
17082 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17083 (v850_gen_movdi): Prototype.
17084 * config/v850/v850.c: Add support for e3v5 architecture.
17085 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17087 (construct_save_jarl): Add e3v5 long JARL support.
17088 (v850_adjust_insn_length): New function. Adjust length of call
17089 insns when using e3v5 instructions.
17090 (v850_gen_movdi): New function: Generate instructions to move a
17092 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17093 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17094 (TARGET_USE_FPU): Enable for e3v5.
17095 (CONST_OK_FOR_W): New macro.
17096 (ADJUST_INSN_LENGTH): Define.
17097 * config/v850/v850.md (UNSPEC_LOOP): Define.
17098 (attr cpu): Add v850e3v5.
17099 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17100 (movdi): New pattern.
17101 (movdi_internal): New pattern.
17102 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17103 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17104 (cstoresf4): Likewise.
17105 (cstoredf4): Likewise.
17106 (insv): New pattern.
17107 (rotlso3_a): New pattern.
17108 (rotlsi3_b): New pattern
17109 (rotlsi3_v850e3v5): New pattern.
17110 (doloop_begin): New pattern.
17111 (fix_loop_counter): New pattern.
17112 (doloop_end): New pattern.
17113 (branch_normal): Add e3v5 long branch support.
17114 (branch_invert): Likewise.
17115 (branch_z_normal): Likewise.
17116 (branch_z_invert): Likewise.
17117 (branch_nz_normal): Likewise.
17118 (branch_nz_invert): Likewise.
17119 (call_internal_short): Add e3v5 register-indirect JARL support.
17120 (call_internal_long): Likewise.
17121 (call_value_internal_short): Likewise.
17122 (call_value_internal_long): Likewise.
17123 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17124 (mloop): New option.
17125 * config.gcc: Add support for configuring v840e3v5 target.
17126 * doc/invoke.texi: Document new v850 specific command line options.
17128 2013-01-31 Paul Koning <ni1d@arrl.net>
17132 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17133 children if parent is a class.
17134 (prune_unused_types_prune): Don't add DW_AT_declaration.
17136 2013-01-31 Richard Biener <rguenther@suse.de>
17138 PR tree-optimization/56157
17139 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17140 match up operand with SLP child.
17142 2013-01-31 Jason Merrill <jason@redhat.com>
17145 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17146 parameters the first time.
17147 (gen_scheduled_generic_parms_dies): Check completeness here.
17149 2013-01-31 Richard Biener <rguenther@suse.de>
17151 PR middle-end/53073
17152 * common.opt (faggressive-loop-optimizations): New flag,
17153 enabled by default.
17154 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17155 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17156 infer_loop_bounds_from_undefined by it.
17158 2013-01-31 Richard Biener <rguenther@suse.de>
17160 PR tree-optimization/56150
17161 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17162 visit virtual operands.
17163 (find_uses_to_rename_bb): Likewise.
17165 2013-01-31 Richard Biener <rguenther@suse.de>
17167 PR tree-optimization/56150
17168 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17169 mixed store non-store stmts.
17171 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17174 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17175 LIBASAN_EARLY_SPEC is defined.
17176 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17177 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17179 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17182 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17183 invalid args instead of ICEing on it.
17184 (ix86_valid_target_attribute_tree): Return error_mark_node if
17185 ix86_valid_target_attribute_inner_p failed.
17186 (ix86_valid_target_attribute_p): Return false only if
17187 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17188 target("default") attribute.
17189 (sorted_attr_string): Change argument from const char * to tree,
17190 merge in all target attribute arguments rather than just one.
17191 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17192 instead of free. Avoid using strcat.
17193 (ix86_mangle_function_version_assembler_name): Mangle
17194 target("default") as if no target attribute is present. Adjust
17195 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
17196 instead of xmalloc and XDELETEVEC instead of free.
17197 (ix86_function_versions): Don't return true if one of the decls
17198 doesn't have target attribute. If they don't and one of the decls
17199 is DECL_FUNCTION_VERSIONED, report an error. Adjust
17200 sorted_attr_string caller. Use XDELETEVEC instead of free.
17201 (ix86_supports_function_versions): Remove.
17202 (make_name): Fix up formatting.
17203 (make_dispatcher_decl): Remove resolver_name and its initialization.
17204 Avoid leaking memory.
17205 (is_function_default_version): Return true if there is
17206 target("default") attribute rather than no target attribute at all.
17207 (make_resolver_func): Avoid leaking memory.
17208 (ix86_generate_version_dispatcher_body): Likewise.
17209 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
17210 * target.def (supports_function_versions): Remove.
17211 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
17212 * doc/tm.texi: Regenerated.
17214 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
17216 PR rtl-optimization/56144
17217 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
17218 for values with side effects.
17220 2013-01-30 Richard Biener <rguenther@suse.de>
17222 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
17223 (sparseset_pop): Likewise.
17224 * cfganal.c (compute_idf): Likewise. Increase work-stack size
17225 to be able to use quick_push in the worker loop.
17227 2013-01-30 Marek Polacek <polacek@redhat.com>
17229 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
17231 2013-01-30 Richard Biener <rguenther@suse.de>
17234 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
17236 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
17238 PR tree-optimization/56064
17239 * fixed-value.c (fixed_from_double_int): New function.
17240 * fixed-value.h (fixed_from_double_int): New prototype.
17241 (const_fixed_from_double_int): New static inline function.
17242 * fold-const.c (native_interpret_fixed): New static function.
17243 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
17244 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
17245 (native_encode_fixed): New static function.
17246 (native_encode_expr) <FIXED_CST>: Use it.
17247 (native_interpret_int): Move double_int worker code to...
17248 * double-int.c (double_int::from_buffer): ...this new static method.
17249 * double-int.h (double_int::from_buffer): Prototype it.
17251 2013-01-30 Richard Biener <rguenther@suse.de>
17253 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
17254 New pointer-map and obstack.
17255 (init_alias_vars): Allocate pointer-map and obstack.
17256 (delete_points_to_sets): Free them.
17257 (find_what_var_points_to): Cache result.
17258 (find_what_p_points_to): Adjust for changed interface of
17259 find_what_var_points_to.
17260 (compute_points_to_sets): Likewise.
17261 (ipa_pta_execute): Likewise.
17263 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17265 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
17266 * configure: Regenerate.
17267 * config.in: Regenerate.
17268 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
17269 #nobits/#progbits if supported.
17271 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
17274 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
17275 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
17276 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
17278 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17280 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
17281 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
17283 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17285 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
17286 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
17288 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17290 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
17292 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
17293 * config/arm/cortex-a7.md: New bypasses using
17294 arm_mac_accumulator_is_result.
17296 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17298 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
17299 (cortex_a7_neon_mla): Likewise.
17300 (cortex_a7_fpfmad): New reservation.
17301 (cortex_a7_fpmacs): Use ffmas and update required units.
17302 (cortex_a7_fpmuld): Update required units and latency.
17303 (cortex_a7_fpmacd): Likewise.
17304 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
17305 (cortex_a7_neon). Likewise.
17306 (bypass) Update participating units.
17308 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17310 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
17311 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
17313 * config/arm/vfp11.md (vfp_farith): Use ffmas.
17314 (vfp_fmul): Use ffmad.
17315 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
17316 (cortex_r4_fmacd): Use ffmad.
17317 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
17318 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
17319 (cortex_a9_fmacd): Use ffmad.
17320 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
17321 (cortex_a8_vfp_macd): Use ffmad.
17322 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
17323 (cortex_a5_fpmacd): Use ffmad.
17324 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
17325 (cortex_a15_vfp_macd): Use ffmad.
17326 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
17328 2013-01-29 Jason Merrill <jason@redhat.com>
17331 * varasm.c (default_assemble_visibility): Don't warn about
17332 visibility on artificial decls.
17334 2013-01-29 Richard Biener <rguenther@suse.de>
17336 PR tree-optimization/56113
17337 * tree-ssa-structalias.c (equiv_class_lookup): Also return
17339 (label_visit): Free duplicate bitmaps and record the leader instead.
17340 (perform_var_substitution): Adjust.
17342 2013-01-29 Richard Biener <rguenther@suse.de>
17344 PR tree-optimization/55270
17345 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
17346 the CFG, schedule loops for fixup.
17348 2013-01-29 Nick Clifton <nickc@redhat.com>
17350 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
17353 2013-01-28 Leif Ekblad <leif@rdos.net>
17355 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
17356 * config/i386/i386.h (TARGET_RDOS): New macro.
17357 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
17358 * config/i386/i386.c (ix86_option_override_internal): For 64bit
17359 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
17360 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
17361 DEFAULT_LARGE_SECTION_THRESHOLD.
17362 * config/i386/i386.md (R14_REG, R15_REG): New constants.
17363 * config/i386/rdos.h: New file.
17364 * config/i386/rdos64.h: New file.
17366 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
17369 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
17372 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17374 PR rtl-optimization/56117
17375 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
17376 call cselib_lookup_from_insn on the MEM before calling
17377 add_insn_mem_dependence.
17379 2013-01-28 Richard Biener <rguenther@suse.de>
17381 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
17382 to a stmt that didn't have one.
17383 (copy_phis_for_bb): Likewise for PHI arguments.
17384 (copy_debug_stmt): Likewise for debug stmts.
17386 2013-01-28 Richard Biener <rguenther@suse.de>
17388 PR tree-optimization/56034
17389 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
17390 (partition_builtin_p): Adjust.
17391 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
17392 it is the last partition.
17393 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
17394 up the vertex for the definition.
17395 (classify_partition): Classify whether a partition is a
17396 PKIND_REDUCTION, thus has uses outside of the loop.
17397 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
17398 Merge all PKIND_REDUCTION partitions into the last partition.
17399 (tree_loop_distribution): Seed partitions from reductions as well.
17401 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17403 PR tree-optimization/56125
17404 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
17405 pow(x,c) into sqrt(x) * powi(x, n/2) or
17406 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
17407 optimizing for size.
17408 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
17409 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
17412 PR tree-optimization/56094
17413 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
17414 to UNKNOWN_LOCATION while gimplifying expr.
17416 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
17419 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
17420 operand 0 in movabs insn template for -masm=intel asm alternative.
17421 (*movabs<mode>_2): Ditto for operand 1.
17423 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
17426 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
17429 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17431 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
17434 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17436 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
17437 Undef to avoid warning.
17439 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17441 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
17442 * configure: Regenerate.
17444 2013-01-25 Jakub Jelinek <jakub@redhat.com>
17446 PR tree-optimization/56098
17447 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
17448 for stmts with volatile ops.
17449 (cond_store_replacement): Don't optimize if assign has volatile ops.
17450 (cond_if_else_store_replacement_1): Don't optimize if either
17451 then_assign or else_assign have volatile ops.
17452 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
17455 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17457 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
17459 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17461 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17462 missing ':' in asm example.
17464 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17466 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17467 entries into lane and laneq entries.
17468 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17469 Remove AdvSIMD scalar modes.
17470 (aarch64_sq<r>dmulh_laneq<mode>): New.
17471 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17473 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17474 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17475 * config/aarch64/iterators.md (VCOND): New.
17478 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17481 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17482 Add NULL LIBNAME argument to existing definitions.
17483 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17484 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17485 * config/avr/avr.c (DEF_BUILTIN): Same.
17486 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17487 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17488 implementation is available (DECL_ASSEMBLER_NAME is set).
17489 (avr_fold_absfx): New static function.
17490 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17491 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17492 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17493 AVR_BUILTIN_ABSLLK.
17494 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17495 (abshk, absk, abslk, absllk): Provide as static inline functions.
17497 2013-01-25 Marek Polacek <polacek@redhat.com>
17499 PR tree-optimization/56035
17500 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17502 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17504 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17505 (*movtf_internal_rex64): Add (!o,C) alternative
17506 (*movxf_internal_rex64): Ditto.
17507 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17509 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17511 * doc/invoke.texi: fix typo.
17512 * doc/objc.texi: fix typo.
17514 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17516 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17517 for the first two alternatives.
17519 2013-01-24 Diego Novillo <dnovillo@google.com>
17521 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17522 (ggc-zone.o): Remove.
17523 * configure.ac: Remove option --with-gc.
17524 * configure: Re-generate.
17525 * doc/install.texi: Remove documentation for --with-gc.
17526 * gengtype.c (write_enum_defn): Remove. Update all users.
17527 (write_Types_process_field): Remove generation of gt_e_* argument.
17528 (output_type_enum): Remove. Update all users.
17529 (write_enum_defn): Remove. Update all users.
17530 (enum alloc_zone): Remove. Update all users.
17531 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17532 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17533 Update all callers.
17534 (struct ptr_data): Remove field TYPE. Update all users.
17535 (gt_pch_note_object): Remove argument TYPE. Update all users.
17536 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17538 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17539 (struct alloc_zone): Remove.
17540 (ggc_internal_alloc_zone_stat): Remove.
17541 (ggc_internal_cleared_alloc_zone_stat): Remove.
17542 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17543 (ggc_pch_count_object): Remove last argument. Update all users.
17544 (ggc_pch_alloc_object): Remove last argument. Update all users.
17545 (struct alloc_zone): Remove.
17546 * ggc-zone.c: Remove.
17547 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17548 (struct alloc_zone): Remove.
17549 (ggc_alloc_typed_stat): Remove.
17550 (ggc_alloc_typed): Remove.
17551 (ggc_splay_alloc): Remove first argument.
17552 (rtl_zone): Remove. Update all users.
17553 (tree_zone): Remove. Update all users.
17554 (tree_id_zone): Remove. Update all users.
17555 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17556 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17557 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17558 * tree-ssanames.c: Remove references to zone allocator in comments.
17560 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17562 * config/avr/avr.c (avr_out_fract): Make register numbers that
17563 might be outside of source operand signed.
17565 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17567 * config/i386/constraints.md (Yf): New constraint.
17568 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17569 of f constraint to conditionaly disable x87 register preferences.
17570 (*movdf_internal): Ditto.
17571 (*movsf_internal): Ditto.
17573 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17575 PR inline-asm/55934
17576 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17577 that have operands with impossible constraints.
17578 Add a FIXME for a speed-up opportunity.
17579 * lra-constraints.c (process_alt_operands): Verify that a class
17580 selected from constraints on asms is valid for the operand mode.
17581 (curr_insn_transform): Remove incorrect comment.
17583 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17585 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17586 TOC operand is a valid symbol ref in the constant pool.
17588 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17590 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17592 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17595 * config/avr/stdfix.h: New file.
17596 * t-avr (stdfix-gcc.h): New rule to build it.
17597 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17599 2013-01-23 Kostya Serebryany <kcc@google.com>
17601 * config/darwin.h: remove dependency on
17602 CoreFoundation (asan on Mac OS).
17604 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17607 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17608 instead of cmpdi_operand for first comparison operand.
17609 Don't assert that comparison operands aren't both constants.
17611 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17613 * doc/install.texi (Downloading the Source): Update references to
17614 downloading separate components.
17616 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17618 * doc/extend.texi (__int128): Improve grammar.
17620 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17623 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17624 alternative to (o,r).
17625 (*movdi_internal_rex64): Remove (!o,n) alternative.
17626 (DImode immediate->memory splitter): Remove.
17627 (DImode immediate->memory peephole2): Remove.
17628 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17629 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17630 alternative to (!o,*r).
17631 (*movtf_internal_sse): New pattern.
17632 (*movxf_internal_rex64): New pattern.
17633 (*movxf_internal): Disable for TARGET_64BIT.
17634 (*movdf_internal_rex64): Remove (!o,F) alternative.
17636 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17638 PR middle-end/56074
17639 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17640 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17641 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17642 stmt locations where LOCATION_LOCUS of the stmt location is
17643 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17646 * config/i386/i386.md (UNSPEC_STOS): New.
17647 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17648 *strsetqi_1): Add UNSPEC_STOS.
17650 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17653 * doc/invoke.texi: Remove left over -Wsynth example.
17655 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17657 PR tree-optimization/56051
17658 * fold-const.c (fold_binary_loc): Don't fold
17659 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17660 a narrowing conversion, or widening conversion from signed
17663 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17665 PR rtl-optimization/56023
17666 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17667 dependent on debug instruction.
17669 2013-01-21 Martin Jambor <mjambor@suse.cz>
17671 PR middle-end/56022
17672 * function.c (allocate_struct_function): Call
17673 invoke_set_current_function_hook earlier.
17675 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17677 * reload1.c (init_reload): Only initialize reload_obstack
17678 during the first call.
17680 2013-01-21 Marek Polacek <polacek@redhat.com>
17682 * cfgloop.c (verify_loop_structure): Fix up grammar.
17684 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17686 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17687 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17689 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17692 * config/arm/marvell-pj4.md: Update copyright year.
17693 Fix up use of alu to alu_reg and simple_alu_imm.
17695 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17697 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17699 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17702 * lra-constraints.c (curr_insn_transform): Don't reuse original
17703 insn for secondary memory move when memory mode should be different.
17705 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17707 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17708 atomic_storedi_1): New patterns.
17710 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17712 btver2 pipeline descriptions.
17713 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17715 * config/i386/i386.md (btver2_decode): New type attributes.
17716 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17718 * config/i386/btver2.md: New file describing btver2 pipelines.
17720 2013-01-19 Andrew Pinski <apinski@cavium.com>
17722 PR tree-optimization/52631
17723 * tree-ssa-sccvn (visit_use): Before looking up the original
17724 statement, try looking up the simplified expression.
17726 2013-01-19 Anthony Green <green@moxielogic.com>
17728 * config/moxie/moxie.c (moxie_expand_prologue): Set
17729 current_function_static_stack_size.
17731 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17733 PR tree-optimization/56029
17734 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17735 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17737 2013-01-18 Sharad Singhai <singhai@google.com>
17739 PR tree-optimization/55995
17740 * dumpfile.c (dump_loc): Print location only if available.
17741 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17743 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17746 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17747 secondary memory move.
17748 (inherit_reload_reg): Use rclass instead of cl for
17749 check_secondary_memory_needed_p.
17751 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17753 PR middle-end/56015
17754 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17755 the case where writing real complex part of target modifies op1.
17757 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17759 * config/aarch64/aarch64-simd.md
17760 (aarch64_vcond_internal<mode>): Handle unordered cases.
17761 * config/aarch64/iterators.md (v_cmp_result): New.
17763 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17764 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17766 * config/arm/marvell-pj4.md: New file.
17767 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17768 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17769 (generic_vfp): Likewise.
17770 * config/arm/arm-cores.def: Add marvell-pj4.
17771 * config/arm/arm-tune.md: Regenerate.
17772 * config/arm/arm-tables.opt: Regenerate.
17773 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17774 * doc/invoke.texi: Document marvell-pj4.
17776 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17778 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17780 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17785 * var-tracking.c (negative_power_of_two_p): New.
17786 (global_get_addr_cache, local_get_addr_cache): New.
17787 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17788 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17790 (vt_stack_offset_p): Remove.
17791 (vt_canon_true_dep): Always canonicalize loc's address.
17792 (clobber_overlapping_mems): Make sure we have a MEM.
17793 (local_get_addr_clear_given_value): New.
17794 (val_reset): Clear local cached entries.
17795 (compute_bb_dataflow): Create and release the local cache.
17796 Disable duplicate MEMs clobbering.
17797 (emit_notes_in_bb): Clobber MEMs likewise.
17798 (vt_emit_notes): Create and release the local cache.
17799 (vt_initialize, vt_finalize): Create and release the global
17800 cache, respectively.
17801 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17803 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17805 PR libmudflap/53359
17806 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17807 not found in the symtab.
17809 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17812 PR rtl-optimization/55547
17813 PR rtl-optimization/53827
17816 * alias.c (offset_overlap_p): New, factored out of...
17817 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17818 the conservative special case for symbolic constants. Don't
17819 adjust zero sizes on alignment.
17821 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17823 PR rtl-optimization/52573
17824 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17825 REG_UNUSED for the same register.
17827 2013-01-17 Richard Biener <rguenther@suse.de>
17828 Marek Polacek <polacek@redhat.com>
17830 PR rtl-optimization/55833
17831 * loop-unswitch.c (unswitch_loops): Move loop verification...
17832 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17833 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17834 Set it to true when we're removing a loop from hierarchy tree in
17835 an irreducible region.
17836 (fix_bb_placements): Adjust caller.
17837 (fix_loop_placements): Likewise.
17839 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17841 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17842 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17843 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17844 Remove ID. Adjust comments.
17845 * config/avr/avr-c.c (avr_builtin_name): Remove.
17846 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17847 * config/avr/avr.c (avr_tolower): New static function.
17848 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17849 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17850 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17853 2013-01-17 Jan Hubicka <jh@suse.cz>
17855 PR tree-optimization/55273
17856 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17858 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17861 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17862 store through atomic_store<mode>_1.
17863 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17865 2013-01-17 Martin Jambor <mjambor@suse.cz>
17867 PR tree-optimizations/55264
17868 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17869 for virtual methods.
17870 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17871 virtual methods before inlining is over.
17872 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17874 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17877 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17879 PR rtl-optimization/56005
17880 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17881 pending reads for prefetch.
17883 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17885 * config/aarch64/aarch64.md
17886 (*cstoresi_neg_uxtw): New pattern.
17887 (*cmovsi_insn_uxtw): New pattern.
17888 (*<optab>si3_uxtw): New pattern.
17889 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17890 (*<optab>si3_insn_uxtw): New pattern.
17891 (*bswapsi2_uxtw): New pattern.
17893 2013-01-16 Richard Biener <rguenther@suse.de>
17895 * tree-inline.c (tree_function_versioning): Remove set but
17896 never used variable.
17898 2013-01-16 Richard Biener <rguenther@suse.de>
17900 PR tree-optimization/55964
17901 * tree-flow.h (rename_variables_in_loop): Remove.
17902 (rename_variables_in_bb): Likewise.
17903 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17904 (copy_loop_before): Adjust and delete update-ssa status.
17905 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17906 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17907 (rename_variables_in_loop): Remove.
17908 (slpeel_update_phis_for_duplicate_loop): Likewise.
17909 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17910 use available cfg machinery instead of duplicating it.
17911 Update PHI nodes and perform poor-mans SSA update here.
17912 (slpeel_tree_peel_loop_to_edge): Adjust.
17914 2013-01-16 Richard Biener <rguenther@suse.de>
17916 PR tree-optimization/54767
17917 PR tree-optimization/53465
17918 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17919 (vrp_visit_phi_node): For PHI arguments coming via backedges
17920 drop all symbolical range information.
17921 (execute_vrp): Compute backedges.
17923 2013-01-16 Richard Biener <rguenther@suse.de>
17925 * doc/install.texi: Update CLooG and ISL requirements to
17928 2013-01-16 Christian Bruel <christian.bruel@st.com>
17931 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17932 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17933 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17935 2013-01-16 DJ Delorie <dj@redhat.com>
17937 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17938 (UNSPECV_SP_SWITCH_E): New.
17939 (sp_switch_1): Change to an unspec.
17940 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17943 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17945 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17946 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17947 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17948 (expand_mem_thread_fence): Ditto.
17949 (expand_mem_signal_fence): Ditto.
17950 (expand_atomic_load): Ditto.
17951 (expand_atomic_store): Ditto.
17953 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17955 PR rtl-optimization/55547
17956 PR rtl-optimization/53827
17959 * alias.c (memrefs_conflict_p): Set sizes to negative after
17962 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17965 * function.c (thread_prologue_and_epilogue_insns): Always
17966 add crtl->drap_reg to set_up_by_prologue.set, even if
17967 stack_realign_drap is false.
17969 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17971 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17972 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17973 *call): Fix indention.
17975 2013-01-15 Tom de Vries <tom@codesourcery.com>
17978 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17981 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17983 PR rtl-optimization/55153
17984 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17986 2013-01-15 Martin Jambor <mjambor@suse.cz>
17988 PR tree-optimization/55920
17989 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17990 accesses as grp_to_be_debug_replaced.
17992 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17994 PR tree-optimization/55920
17995 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17996 there is non-useless type conversion needed from debug rhs to lhs,
17997 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17999 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18000 Mikael Pettersson <mikpe@it.uu.se>
18003 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18005 (ASM_OUTPUT_CASE_LABEL): Remove.
18006 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18007 * final.c (shorten_branches): Update alignment of labels before
18008 jump tables if CASE_VECTOR_SHORTEN_MODE.
18010 2013-01-15 Richard Biener <rguenther@suse.de>
18013 * system.h: Do not include gmp.h for building host tools.
18015 2013-01-15 Richard Biener <rguenther@suse.de>
18017 PR middle-end/55882
18018 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18019 account for bitpos when computing alignment.
18021 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18023 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18024 (ix86_target_macros_internal): Likewise.
18026 * config/i386/i386.c (m_CORE2I7): Removed.
18027 (m_CORE_HASWELL): New macro.
18028 (m_CORE_ALL): Likewise.
18029 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18030 (initial_ix86_arch_features): Likewise.
18031 (processor_target_table): Initializations for Core avx2.
18032 (cpu_names): New names "core-avx2".
18033 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18034 PROCESSOR_CORE_HASWELL.
18035 (ix86_issue_rate): New case.
18036 (ia32_multipass_dfa_lookahead): Likewise.
18037 (ix86_sched_init_global): Likewise.
18039 * config/i386/i386.h (TARGET_HASWELL): New macro.
18040 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18041 (processor_type): New PROCESSOR_HASWELL.
18043 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18045 PR tree-optimization/55955
18046 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18047 *SHIFT_EXPR and *ROTATE_EXPR codes.
18049 PR tree-optimization/48766
18050 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18051 -ftrapv disable -fwrapv.
18053 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18056 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18057 etc. to 1 and not to __flash.
18058 Use LL suffix for __INT24_MAX__ with -mint8.
18059 Use ULL suffix for __UINT24_MAX__ with -mint8.
18061 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18063 * config/avr/avr-arch.h
18064 (struct base_arch_s): Use typedef avr_arch_t instead.
18065 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18066 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18067 * config/avr/avr.c: Same.
18068 * config/avr/avr-devices.c: Same.
18069 * config/avr/driver-avr.c: Same.
18070 * config/avr/gen-avr-mmcu-texi.c: Same.
18071 * config/avr/avr-mcus.def: Adjust comment.
18073 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18075 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18076 * config/aarch64/iterators.md (VALLDI): New.
18078 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18079 Andi Kleen <ak@linux.intel.com>
18082 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18083 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18086 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18088 * config/avr/avr-stdint.h: Remove trailing blanks.
18089 * config/avr/avr-log.h: Same.
18090 * config/avr/avr-arch.h: Same.
18091 * config/avr/avr-devices.c: Same.
18092 * config/avr/avr-dimode.md: Same.
18093 * config/avr/predicates.md: Same.
18094 * config/avr/avr-c.c: Same. And fix typo.
18096 * config/avr/avr-protos.h: Same. And:
18097 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18098 (init_cumulative_args): Rename to avr_init_cumulative_args.
18099 (expand_prologue): Rename to avr_expand_prologue.
18100 (expand_epilogue): Rename to avr_expand_epilogue.
18101 (adjust_insn_length): Rename to avr_adjust_insn_length.
18102 (notice_update_cc): Rename to avr_notice_update_cc.
18103 (final_prescan_insn): Rename to avr_final_prescan_insn.
18104 * config/avr/avr.c: Same.
18105 * config/avr/avr.h: Same.
18106 * config/avr/avr.md: Remove trailing blanks.
18107 (prologue): Use avr_expand_prologue.
18108 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18110 2013-01-14 Richard Biener <rguenther@suse.de>
18112 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18113 verify_location, collect_subblocks): New functions.
18114 (verify_gimple_in_cfg): Verify that locations only reference
18115 BLOCKs in the functions BLOCK tree.
18117 2013-01-14 Richard Biener <rguenther@suse.de>
18119 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18121 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18123 (insert_out_of_ssa_copy_on_edge): Likewise.
18124 (rewrite_close_phi_out_of_ssa): Likewise.
18125 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18127 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18128 propagated constants.
18129 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18132 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18134 * config/avr/avr-modes.def: Add GPL copyright notice.
18136 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18138 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18139 MEMMODEL_MASK to determine memory model.
18140 (atomic_store<mode>): Ditto from operands[2].
18141 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18143 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18146 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18147 (fold_gimple_assign): Don't call unshare_expr here.
18148 (fold_ctor_reference): Call unshare_expr.
18150 2013-01-13 Terry Guo <terry.guo@arm.com>
18152 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18153 * doc/fragments.texi: Document MULTILIB_REUSE.
18154 * gcc.c (multilib_reuse): New internal spec.
18155 (set_multilib_dir): Also search multilib from multilib_reuse.
18156 * genmultilib (tmpmultilib3): Refactor code.
18157 (tmpmultilib4): Ditto.
18158 (multilib_reuse): New multilib argument.
18160 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18162 * Makefile.in: Update copyright.
18164 2013-01-12 Tom de Vries <tom@codesourcery.com>
18166 PR middle-end/55890
18167 * calls.c (expand_call): Check if arg_nr is valid.
18169 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18171 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18172 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18173 documentation. Add missing '__' in front of
18174 __builtin_ia32_packssdw256.
18176 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18179 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18180 NO_REGS for larl operands.
18181 (s390_reload_larl_operand): Use s390_load_address instead of
18184 2013-01-11 Richard Biener <rguenther@suse.de>
18186 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18187 (verify_node_sharing): ... here.
18188 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18190 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18192 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18193 Substitute TREECHECKING.
18194 * configure: Regenerate.
18195 * Makefile.in (TREECHECKING): New.
18197 2013-01-11 Richard Guenther <rguenther@suse.de>
18199 PR tree-optimization/44061
18200 * tree-vrp.c (extract_range_basic): Compute zero as
18201 value-range for __builtin_constant_p of function parameters.
18203 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
18205 Update copyright years.
18207 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
18209 PR rtl-optimization/55672
18210 * lra-eliminations.c (mark_not_eliminable): Permit addition with
18211 const to be eliminable.
18213 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
18215 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
18216 * configure: Regenerate.
18218 2013-01-10 Richard Biener <rguenther@suse.de>
18220 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
18222 2013-01-10 Richard Biener <rguenther@suse.de>
18225 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
18226 locations for virtual PHI arguments.
18227 (rewrite_update_phi_arguments): Likewise.
18229 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
18231 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
18234 2013-01-10 Jakub Jelinek <jakub@redhat.com>
18236 PR tree-optimization/55921
18237 * tree-complex.c (expand_complex_asm): New function.
18238 (expand_complex_operations_1): Call it for GIMPLE_ASM.
18240 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18243 * config/s390/s390.c (s390_symref_operand_p)
18244 (s390_loadrelative_operand_p): Merge the two functions.
18245 (s390_check_qrst_address, print_operand_address): Add parameters
18246 to s390_loadrelative_operand_p invokation.
18247 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
18248 (s390_reload_larl_operand, s390_secondary_reload): Use
18249 s390_loadrelative_operand_p instead of s390_symref_operand_p.
18250 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
18252 2013-01-09 Mike Stump <mikestump@comcast.net>
18254 * dse.c (record_store): Remove unnecessary assert.
18256 2013-01-09 Jan Hubicka <jh@suse.cz>
18258 PR tree-optimization/55569
18259 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
18260 * cfgloop.h (scale_loop_profile): Likewise.
18262 2013-01-09 Jan Hubicka <jh@suse.cz>
18265 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
18267 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
18269 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
18271 PR middle-end/55114
18272 * expr.h (maybe_emit_group_store): Declare.
18273 * expr.c (maybe_emit_group_store): New function.
18274 * builtins.c (expand_builtin_int_roundingfn): Call it.
18275 (expand_builtin_int_roundingfn_2): Likewise.
18277 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
18279 PR rtl-optimization/55829
18280 * lra-constraints.c (match_reload): Add code for absent output.
18281 (curr_insn_transform): Add code for reloads of matched inputs
18284 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
18286 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
18287 attribute of movddup insn to DF.
18288 (*vec_interleave_lowv2df): Ditto.
18289 (vec_dupv2df): Ditto.
18291 2013-01-09 Jan Hubicka <jh@suse.cz>
18293 PR tree-optimiation/55875
18294 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
18295 EVERY_ITERATION parameter.
18296 (number_of_iterations_exit): Check if exit is executed every iteration.
18297 (idx_infer_loop_bounds): Similarly here.
18298 (n_of_executions_at_most): Simplify
18299 to only test for cases where statement is dominated by the
18300 particular bound; handle correctly the "postdominance" test.
18301 (scev_probably_wraps_p): Use max loop iterations info
18302 as a global bound first.
18304 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
18305 Nick Clifton <nickc@redhat.com>
18307 * config/v850/v850.md (cbranchsf4): New pattern.
18308 (cstoresf4): New pattern.
18309 (cbranchdf4): New pattern.
18310 (cstoredf4): New pattern.
18311 (movsicc): Disallow floating point comparisons.
18312 (cmpsf_le_insn): Fix order of operators.
18313 (cmpsf_lt_insn): Likewise.
18314 (cmpsf_eq_insn): Likewise.
18315 (cmpdf_le_insn): Likewise.
18316 (cmpdf_lt_insn): Likewise.
18317 (cmpdf_eq_insn): Likewise.
18318 (cmpsf_ge_insn): Use LE comparison.
18319 (cmpdf_ge_insn): Likewise.
18320 (cmpsf_gt_insn): Use LT comparison.
18321 (cmpdf_gt_insn): Likewise.
18322 (cmpsf_ne_insn): Delete pattern.
18323 (cmpdf_ne_insn): Delete pattern.
18324 * config/v850/v850.c (v850_gen_float_compare): Use
18325 gen_cmpdf_eq_insn for NE comparison.
18326 (v850_float_z_comparison_operator)
18327 (v850_float_nz_comparison_operator): Move from here ...
18328 * config/v850/predicates.md: ... to here. Move GT and GE
18329 comparisons into v850_float_z_comparison_operator.
18330 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
18332 (v850_float_nz_comparison_operator): Likewise.
18334 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18336 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
18337 with calls to gen_insvsi/gen_insvdi.
18339 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18341 * config/i386/i386.c (initial_ix86_tune_features): Set up
18342 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
18344 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
18345 Jakub Jelinek <jakub@redhat.com>
18347 PR tree-optimization/48189
18348 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
18349 If nitercst is 0, don't predict the exit edge.
18351 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18353 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
18354 in asm_fprintf with reg_names.
18355 (aarch64_print_operand_address): Likewise.
18356 (aarch64_return_addr): Likewise.
18357 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
18359 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18361 * config/pa/pa.h (VAL_U6_BITS_P): Define.
18362 (INT_U6_BITS): Likewise.
18363 * config/pa/predicates.md (uint6_operand): New predicate.
18364 (shift5_operand, shift6_operand): Likewise.
18365 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
18367 (lshrdi3): Use shift6_operand.
18368 (shrpsi4, shrpdi4): New insn patterns.
18369 (extzv): Delete expander.
18370 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
18371 predicates in unamed zero extract patterns. Tighten common constraint.
18372 (extv): Delete expander.
18373 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
18374 predicates in unamed sign extract patterns. Tighten common constraint.
18375 (insv): Delete expander.
18376 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
18377 predicates in unamed insert patterns. Tighten common constraint.
18378 Change uint32_operand predicate to uint6_operand predicate in unamed
18379 DImode pattern to insert constant values of type 1...1xxxx.
18381 2013-01-04 Jan Hubicka <jh@suse.cz>
18383 PR tree-optimization/55823
18384 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
18387 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18388 Uros Bizjak <ubizjak@gmail.com>
18390 PR rtl-optimization/55845
18391 * df-problems.c (can_move_insns_across): Stop scanning at
18392 volatile_insn_p source instruction or give up if
18393 across_from .. across_to range contains any volatile_insn_p
18396 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18398 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
18399 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
18401 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
18402 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
18404 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18407 * asan.c (asan_clear_shadow): New function.
18408 (asan_emit_stack_protection): Use it.
18410 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18412 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
18413 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
18414 with tab instead of space.
18416 2013-01-08 Nick Clifton <nickc@redhat.com>
18418 * config/rl78/rl78.c (rl78_expand_prologue): Always select
18419 register bank 0 at the start of an interrupt handler.
18420 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
18423 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18425 * config/aarch64/aarch64-simd.md
18426 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
18427 (aarch64_simd_bsl): Likewise.
18428 (aarch64_vcond_internal<mode>): Likewise.
18429 (vcond<mode><mode>): Likewise.
18430 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
18431 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
18433 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18435 * config/aarch64/aarch64-builtins.c
18436 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
18438 2013-01-08 Martin Jambor <mjambor@suse.cz>
18441 * tree-sra.c (analyze_access_subtree): Return true also after
18442 potentially creating a debug-only replacement.
18444 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18446 PR middle-end/55890
18447 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
18449 PR tree-optimization/54120
18450 * tree-vrp.c (range_fits_type_p): Don't allow
18451 src_precision < precision from signed vr to unsigned_p
18452 if vr->min or vr->max is negative.
18453 (simplify_float_conversion_using_ranges): Test can_float_p
18454 against CODE_FOR_nothing.
18456 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18457 Richard Biener <rguenther@suse.de>
18459 PR middle-end/55851
18460 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
18461 types instead of just INTEGER_TYPE types.
18463 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18465 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18468 2013-01-07 Steve Ellcey <sellcey@mips.com>
18471 * config/mips/mips.opt: Change mad to mmad to match documentation.
18473 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18476 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18477 .progmemx.data now.
18479 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18482 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18483 (avr_addrspace_t): Add .section_name field.
18484 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18486 (avr_addrspace): Same. Initialize .section_name. Remove last
18487 NULL entry. Put __memx into .progmemx.data.
18488 (progmem_section_prefix): Remove.
18489 (avr_asm_init_sections): No need to initialize progmem_section.
18490 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18491 section name prefix.
18492 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18493 retrieve the progmem section.
18494 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18495 boundary to run over avr_addrspace[].
18496 (avr_register_target_pragmas): Ditto.
18498 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18500 * varasm.c (output_constant_def_contents): For asan_protect_global
18501 protected strings, adjust DECL_ALIGN if needed, before testing for
18503 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18504 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18506 (output_object_block): For asan protected decls, emit asan padding
18507 after their contents.
18508 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18509 (asan_finish_file): Test it here instead.
18511 2013-01-07 Nick Clifton <nickc@redhat.com>
18512 Matthias Klose <doko@debian.org>
18513 Doug Kwan <dougkwan@google.com>
18514 H.J. Lu <hongjiu.lu@intel.com>
18517 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18519 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18521 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18523 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18525 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18527 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18530 * doc/install.texi (Cross-Compiler-Specific Options): Document
18533 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18535 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18536 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18537 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18538 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18539 vqmovun_high_s64): Fix source operand number and update copyright.
18541 2013-01-07 Richard Biener <rguenther@suse.de>
18543 PR middle-end/55890
18544 * gimple.h (gimple_call_builtin_p): New overload.
18545 * gimple.c (validate_call): New function.
18546 (gimple_call_builtin_p): Likewise.
18547 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18548 Use gimple_call_builtin_p.
18549 (find_func_clobbers): Likewise.
18550 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18551 (strlen_optimize_stmt): Likewise.
18553 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18555 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18556 (vld1q_dup_*): Likewise.
18557 (vld1_*): Likewise.
18558 (vld1q_*): Likewise.
18559 (vld1_lane_*): Likewise.
18560 (vld1q_lane_*): Likewise.
18562 2013-01-07 Richard Biener <rguenther@suse.de>
18564 * lto-streamer.h (LTO_minor_version): Bump to 2.
18566 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18568 * config/aarch64/aarch64-protos.h
18569 (aarch64_const_double_zero_rtx_p): Rename to...
18570 (aarch64_float_const_zero_rtx_p): ...this.
18571 (aarch64_float_const_representable_p): New.
18572 (aarch64_output_simd_mov_immediate): Likewise.
18573 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18574 move immediate case.
18575 * config/aarch64/aarch64.c
18576 (aarch64_const_double_zero_rtx_p): Rename to...
18577 (aarch64_float_const_zero_rtx_p): ...this.
18578 (aarch64_print_operand): Allow printing of new constants.
18579 (aarch64_valid_floating_const): New.
18580 (aarch64_legitimate_constant_p): Check for valid floating-point
18582 (aarch64_simd_valid_immediate): Likewise.
18583 (aarch64_vect_float_const_representable_p): New.
18584 (aarch64_float_const_representable_p): Likewise.
18585 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18586 (aarch64_output_simd_mov_immediate): New.
18587 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18588 (*movdf_aarch64): Likewise.
18589 * config/aarch64/constraints.md (Ufc): New.
18590 (Y): call aarch64_float_const_zero_rtx.
18591 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18593 2013-01-07 Richard Biener <rguenther@suse.de>
18595 PR tree-optimization/55888
18596 PR tree-optimization/55862
18597 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18598 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18599 not if it is contained therein.
18601 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18603 * config/avr/t-avr: Typo.
18605 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18608 * config/avr/t-avr: Don't automatically rebuild
18609 $(srcdir)/config/avr/t-multilib
18610 $(srcdir)/config/avr/avr-tables.opt
18611 $(srcdir)/doc/avr-mmcu.texi
18612 (avr-mcus): New phony target to build them on request.
18613 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18614 * avr/avr-mcus.def: Adjust comments.
18616 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18618 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18620 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18622 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18624 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18626 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18628 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18630 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18631 to generate profiling.
18632 * config/rs6000/aix64.h (LIB_SPEC): Same.
18634 2013-01-04 Andrew Pinski <apinski@cavium.com>
18636 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18638 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18640 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18642 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18644 (ix86_expand_move): Ditto.
18645 (ix86_zero_extend_to_Pmode): Ditto.
18646 (ix86_expand_call): Ditto.
18647 (ix86_expand_special_args_builtin): Ditto.
18648 (ix86_expand_builtin): Ditto.
18650 2013-01-04 Richard Biener <rguenther@suse.de>
18652 PR tree-optimization/55862
18653 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18654 translating them through PHI nodes.
18656 2013-01-04 Martin Jambor <mjambor@suse.cz>
18658 PR tree-optimization/55755
18659 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18660 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18662 2013-01-04 Marek Polacek <polacek@redhat.com>
18664 PR middle-end/55859
18665 * opts.c (default_options_optimization): Clarify error message.
18667 2013-01-04 Richard Biener <rguenther@suse.de>
18669 PR middle-end/55863
18670 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18673 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18676 * config/pa/pa.md (movsi): Revert previous change.
18677 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18680 2013-01-03 Richard Henderson <rth@redhat.com>
18682 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18683 after eliminating TLS symbols.
18685 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18688 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18689 * graphite-poly.c (debug_gmp_value): Likewise.
18691 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18694 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18695 selected code model, define __code_mode_small__, __code_model_medium__,
18696 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18697 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18698 xchg temporary register with %k. Declare temporary register as
18700 [__x86_64__]: For medium and large code models, preserve %rbx register.
18702 2013-01-03 Richard Biener <rguenther@suse.de>
18704 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18705 (dump_subscript): Adjust.
18706 (finalize_ddr_dependent): Do not dump redundant info.
18707 (analyze_siv_subscript): Adjust.
18708 (subscript_dependence_tester): Likewise.
18709 (compute_affine_dependence): Likewise.
18711 2013-01-03 Richard Biener <rguenther@suse.de>
18714 2013-01-03 Richard Biener <rguenther@suse.de>
18716 PR tree-optimization/55857
18717 * tree-vect-stmts.c (vectorizable_load): Do not setup
18718 re-alignment for invariant loads.
18720 2013-01-02 Richard Biener <rguenther@suse.de>
18722 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18723 invariant load do not generate a vector load from the scalar location.
18725 2013-01-03 Richard Biener <rguenther@suse.de>
18727 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18728 for not vectorizing.
18729 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18730 not build INDIRECT_REFs, call get_name once only.
18731 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18732 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18734 2013-01-03 Richard Biener <rguenther@suse.de>
18736 PR tree-optimization/55857
18737 * tree-vect-stmts.c (vectorizable_load): Do not setup
18738 re-alignment for invariant loads.
18740 2013-01-03 Richard Biener <rguenther@suse.de>
18743 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18744 prefer a built-in decl.
18746 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18748 * gcc.c (process_command): Update copyright notice dates.
18749 * gcov.c (print_version): Likewise.
18750 * gcov-dump.c (print_version): Likewise.
18752 PR rtl-optimization/55838
18753 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18754 iv0.step, iv1.step and step.
18756 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18757 Marc Glisse <marc.glisse@inria.fr>
18759 PR tree-optimization/55832
18760 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18761 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18762 integer_{one,zero}_node.
18764 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18767 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18768 * var-tracking.c (reverse_op): Don't add reverse ops to
18769 VALUEs that have already
18770 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18772 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18774 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18776 2013-01-02 Teresa Johnson <tejohnson@google.com>
18778 * dumpfile.c (dump_loc): Print filename with location.
18779 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18780 new location_t parameter to emit complete unroll message with
18781 new dump framework.
18782 (canonicalize_loop_induction_variables): Compute loops location
18783 and pass to try_unroll_loop_completely.
18784 * loop-unroll.c (report_unroll_peel): New function.
18785 (peel_loops_completely): Use new dump format with location
18786 for main dumpfile message, and invoke report_unroll_peel on success.
18787 (decide_unrolling_and_peeling): Ditto.
18788 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18789 by report_unroll_peel.
18790 (decide_peel_completely): Ditto.
18791 (decide_unroll_constant_iterations): Ditto.
18792 (decide_unroll_runtime_iterations): Ditto.
18793 (decide_peel_simple): Ditto.
18794 (decide_unroll_stupid): Ditto.
18795 * cfgloop.c (get_loop_location): New function.
18796 * cfgloop.h (get_loop_location): Declare.
18798 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18800 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18803 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18805 PR middle-end/55198
18806 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18807 BLKmode objects when EXPAND_MEMORY is specified.
18809 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18811 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18813 (fold_builtin_cpu): Do not share cpu model decls across statements.
18815 2013-01-02 Jason Merrill <jason@redhat.com>
18818 * tree.c (build_array_type_1): Revert earlier change.
18820 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18822 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18824 * config/aarch64/aarch64-tune.md: Re-generate.
18826 2013-01-02 Richard Biener <rguenther@suse.de>
18828 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18829 invariant load do not generate a vector load from the scalar location.
18831 2013-01-02 Richard Biener <rguenther@suse.de>
18834 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18835 * configure: Regenerate.
18837 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18839 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18840 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18841 (expand_builtin_int_roundingfn_2): Keep the original target around
18842 for the fallback case.
18844 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18846 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18847 to be clear for sign changes.
18849 2013-01-01 Jan Hubicka <jh@suse.cz>
18851 * ipa-inline-analysis.c: Fix formatting.
18853 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18855 PR tree-optimization/55831
18856 * tree-vect-loop.c (get_initial_def_for_induction): Use
18857 gsi_after_labels instead of gsi_start_bb.
18859 Copyright (C) 2013 Free Software Foundation, Inc.
18861 Copying and distribution of this file, with or without modification,
18862 are permitted in any medium without royalty provided the copyright
18863 notice and this notice are preserved.